0.22.0

What's Changed

• Add cached tokens tracking to LLM usage reporting by @aantn in #1680
• Add eval comparison with master branch by @aantn in #1682
• Add token count caching and performance instrumentation by @aantn in #1692
• Docs on runbooks + remove deprecated by @aantn in #1693
• Update MCP server icon URLs to use GitHub-hosted SVG logos by @aantn in #1670
• Remove custom runbooks example file by @aantn in #1697
• Parallelize KIND cluster setup with HolmesGPT environment by @aantn in #1687
• Remove automated investigation API and consolidate to conversational interface by @aantn in #1688
• ROB-3410 Disable unconfigured toolsets by default in CLI by @naomi-robusta in #1658
• Reorder CONFIG command in SlashCommands enum by @naomi-robusta in #1701
• Add URL query parameter support to tab synchronization by @aantn in #1702
• Track explicitly set fields and improve null/empty value handling by @naomi-robusta in #1703
• Add GitHub App token manager for automatic installation token handling by @Avi-Robusta in #1664
• Add MCP Server documentation link to toolset configuration by @naomi-robusta in #1705
• Merge enabled and status fields into single status display by @naomi-robusta in #1704
• Update welcome banner to mention CONFIG command by @naomi-robusta in #1707
• Simplify litellm.modify_params handling by @aantn in #1709
• Add Enum field support and MCP config management to TUI by @naomi-robusta in #1706
• Improve OOM error messaging to clarify memory limits are by design by @aantn in #1714
• Fix kubernetes tools to return json by @aantn in #1337
• Document memory safety features and OOM protection mechanisms by @aantn in #1715
• Reorganize HTTP connectors documentation into dedicated section by @aantn in #1708
• fix(bedrock): handle IRSA authentication via AWS_ROLE_ARN and AWS_WEB_IDENTITY_TOKEN_FILE by @saintetix in #1671
• Add initContainers and args support to Helm chart by @aantn in #1516
• Add digest auth and mTLS support to HTTP toolset by @aantn in #1720
• Add mTLS (mutual TLS) support to Elasticsearch toolset by @aantn in #1660
• Add outdated version warning banner to documentation by @aantn in #1678
• feat: support keyless auth for Azure foundry by @mainred in #1648
• toolset/inspektor-gadget: Introduce filtering for tools by @aantn in #1724
• Make Grafana config class configurable in base toolset by @aantn in #1732
• Fix double-quoting of filter expressions in inspektor_gadget toolset by @aantn in #1735
• Clarify HolmesGPT works with any infrastructure, not just Kubernetes by @aantn in #1734
• A user that is using custom certificate - OS Error by @arikalon1 in #1719
• Add HolmesGPT version to AI assistant introduction by @aantn in #1736
• Add MongoDB toolset for querying and diagnostics by @arikalon1 in #1731
• [ROB-3057] fix holmes overconfidence by @Avi-Robusta in #1711
• [ROB-3039] GitHub App credentials mcp config by @Avi-Robusta in #1717
• [ROB-3242] Eval fix by @Avi-Robusta in #1740
• Document behavior_controls API parameter for prompt customization by @aantn in #1743
• Weekly Benchmark Results 2026-03-11_21-08 by @github-actions[bot] in #1749
• Add large Confluence page evaluation test case by @aantn in #1738
• Add max_prompt_tokens_per_call tracking to LLM cost reporting by @aantn in #1755
• fix(llm): load MODEL env fallback and improve no-model guidance by @pavangudiwada in #1723
• Fix overly-big-toolcalls handling by @aantn in #1756
• Allow returning to toolset selection after configuration by @aantn in #1758
• Add refactoring plan for unifying call() and call_stream() by @aantn in #1763
• Weekly Benchmark Results 2026-03-14_20-45 by @github-actions[bot] in #1766
• Weekly Benchmark Results 2026-03-15_04-11 by @github-actions[bot] in #1772
• Refactor call() to unify with call_stream() and other improvements by @aantn in #1765
• benchmark summary by @Avi-Robusta in #1776
• Update OpenAI icon from simple-openai to fontawesome-brands-openai by @aantn in #1777
• Support multi-round approval workflows with iteration offset by @aantn in #1774
• fix: enrich PagerDuty issues with description and alert body details by @yakir-shriker in #1780
• Implement schema resolution for JSON Schema references and compound types by @mouchar in #1713
• Rename grafana-dashboard tag to grafana by @aantn in #1786
• fix: forward --model parameter in investigate ticket command by @yakir-shriker in #1779
• Add tool_results_dir parameter to ToolCallingLLM in custom_llm example by @aantn in #1781
• Add new integrations and update integration documentation by @aantn in #1788
• Add support for evals-model-* labels in regression workflow by @aantn in #1784
• Refactor config class handling to support multiple config classes by @naomi-robusta in #1741
• Document CLASSIFIER_MODEL env var and OpenRouter requirements by @aantn in #1792
• holmes-mcp docs update by @Avi-Robusta in #1793
• Soften runbook fetching requirements and reduce enforcement language by @aantn in #1787
• Clarify multi-account AWS setup and agent deployment options by @aantn in #1795
• Strip newlines from sanitized parameters to prevent shell syntax errors by @aantn in #1773
• Enable strict tool calling universally with per-tool compatibility checks by @aantn in #1790
• Remove message truncation logic, fail fast on context overflow by @aantn in #1797
• [ROB-3260] Fix Anthropic image token counting by @Avi-Robusta in #1642
• Document tool approval behavior and update API reference by @aantn in #1800
• Add new k8s evals for kubernetes_jq_query by @aantn in #1759
• Add PR label support and improve eval parameter naming by @aantn in #1799
• Add MCP Confluence image attachment test case (eval 233) by @aantn in #1796
• ROB-3472: Add datasource-catalog.json by @alonelish in #1778
• Patch CVE-2025-68121 by @moshemorad in #1667
• ROB-3472: Update datasource-catalog.json by @alonelish in #1806

New Contributors

• @saintetix made their first contribution in #1671
• @yakir-shriker made their first contribution in #1780
• @mouchar made their first contribution in #1713
• @alonelish made their first contribution in #1778

Full Changelog: 0.21.0...0.22.0

0.22.0-alpha.1

What's Changed

• [ROB-3242] Eval fix by @Avi-Robusta in #1740
• Document behavior_controls API parameter for prompt customization by @aantn in #1743
• Weekly Benchmark Results 2026-03-11_21-08 by @github-actions[bot] in #1749
• Add large Confluence page evaluation test case by @aantn in #1738
• Add max_prompt_tokens_per_call tracking to LLM cost reporting by @aantn in #1755
• fix(llm): load MODEL env fallback and improve no-model guidance by @pavangudiwada in #1723
• Fix overly-big-toolcalls handling by @aantn in #1756
• Allow returning to toolset selection after configuration by @aantn in #1758
• Add refactoring plan for unifying call() and call_stream() by @aantn in #1763
• Weekly Benchmark Results 2026-03-14_20-45 by @github-actions[bot] in #1766
• Weekly Benchmark Results 2026-03-15_04-11 by @github-actions[bot] in #1772
• Refactor call() to unify with call_stream() and other improvements by @aantn in #1765
• benchmark summary by @Avi-Robusta in #1776
• Update OpenAI icon from simple-openai to fontawesome-brands-openai by @aantn in #1777
• Support multi-round approval workflows with iteration offset by @aantn in #1774
• fix: enrich PagerDuty issues with description and alert body details by @yakir-shriker in #1780
• Implement schema resolution for JSON Schema references and compound types by @mouchar in #1713
• Rename grafana-dashboard tag to grafana by @aantn in #1786
• fix: forward --model parameter in investigate ticket command by @yakir-shriker in #1779
• Add tool_results_dir parameter to ToolCallingLLM in custom_llm example by @aantn in #1781
• Add new integrations and update integration documentation by @aantn in #1788
• Add support for evals-model-* labels in regression workflow by @aantn in #1784
• Refactor config class handling to support multiple config classes by @naomi-robusta in #1741
• Document CLASSIFIER_MODEL env var and OpenRouter requirements by @aantn in #1792
• holmes-mcp docs update by @Avi-Robusta in #1793
• Soften runbook fetching requirements and reduce enforcement language by @aantn in #1787
• Clarify multi-account AWS setup and agent deployment options by @aantn in #1795
• Strip newlines from sanitized parameters to prevent shell syntax errors by @aantn in #1773
• Enable strict tool calling universally with per-tool compatibility checks by @aantn in #1790
• Remove message truncation logic, fail fast on context overflow by @aantn in #1797
• [ROB-3260] Fix Anthropic image token counting by @Avi-Robusta in #1642
• Document tool approval behavior and update API reference by @aantn in #1800
• Add new k8s evals for kubernetes_jq_query by @aantn in #1759
• Add PR label support and improve eval parameter naming by @aantn in #1799
• Add MCP Confluence image attachment test case (eval 233) by @aantn in #1796
• ROB-3472: Add datasource-catalog.json by @alonelish in #1778
• Patch CVE-2025-68121 by @moshemorad in #1667

New Contributors

• @yakir-shriker made their first contribution in #1780
• @mouchar made their first contribution in #1713
• @alonelish made their first contribution in #1778

Full Changelog: 0.22.0-alpha...0.22.0-alpha.1

0.22.0-alpha

What's Changed

• Add cached tokens tracking to LLM usage reporting by @aantn in #1680
• Add eval comparison with master branch by @aantn in #1682
• Add token count caching and performance instrumentation by @aantn in #1692
• Docs on runbooks + remove deprecated by @aantn in #1693
• Update MCP server icon URLs to use GitHub-hosted SVG logos by @aantn in #1670
• Remove custom runbooks example file by @aantn in #1697
• Parallelize KIND cluster setup with HolmesGPT environment by @aantn in #1687
• Remove automated investigation API and consolidate to conversational interface by @aantn in #1688
• ROB-3410 Disable unconfigured toolsets by default in CLI by @naomi-robusta in #1658
• Reorder CONFIG command in SlashCommands enum by @naomi-robusta in #1701
• Add URL query parameter support to tab synchronization by @aantn in #1702
• Track explicitly set fields and improve null/empty value handling by @naomi-robusta in #1703
• Add GitHub App token manager for automatic installation token handling by @Avi-Robusta in #1664
• Add MCP Server documentation link to toolset configuration by @naomi-robusta in #1705
• Merge enabled and status fields into single status display by @naomi-robusta in #1704
• Update welcome banner to mention CONFIG command by @naomi-robusta in #1707
• Simplify litellm.modify_params handling by @aantn in #1709
• Add Enum field support and MCP config management to TUI by @naomi-robusta in #1706
• Improve OOM error messaging to clarify memory limits are by design by @aantn in #1714
• Fix kubernetes tools to return json by @aantn in #1337
• Document memory safety features and OOM protection mechanisms by @aantn in #1715
• Reorganize HTTP connectors documentation into dedicated section by @aantn in #1708
• fix(bedrock): handle IRSA authentication via AWS_ROLE_ARN and AWS_WEB_IDENTITY_TOKEN_FILE by @saintetix in #1671
• Add initContainers and args support to Helm chart by @aantn in #1516
• Add digest auth and mTLS support to HTTP toolset by @aantn in #1720
• Add mTLS (mutual TLS) support to Elasticsearch toolset by @aantn in #1660
• Add outdated version warning banner to documentation by @aantn in #1678
• feat: support keyless auth for Azure foundry by @mainred in #1648
• toolset/inspektor-gadget: Introduce filtering for tools by @aantn in #1724
• Make Grafana config class configurable in base toolset by @aantn in #1732
• Fix double-quoting of filter expressions in inspektor_gadget toolset by @aantn in #1735
• Clarify HolmesGPT works with any infrastructure, not just Kubernetes by @aantn in #1734
• A user that is using custom certificate - OS Error by @arikalon1 in #1719
• Add HolmesGPT version to AI assistant introduction by @aantn in #1736
• Add MongoDB toolset for querying and diagnostics by @arikalon1 in #1731
• [ROB-3057] fix holmes overconfidence by @Avi-Robusta in #1711
• [ROB-3039] GitHub App credentials mcp config by @Avi-Robusta in #1717

New Contributors

• @saintetix made their first contribution in #1671

Full Changelog: 0.21.0...0.22.0-alpha

0.21.0

What's Changed

• Add getting started guidance for recommended data source setup by @aantn in #1592
• Reorganize multiple providers documentation for CLI and Helm by @aantn in #1643
• [ROB-3391] Coralogix archive support by @Avi-Robusta in #1650
• ROB-3382 Implement lazy initialization for toolset prerequisites in CLI mode by @naomi-robusta in #1630
• Confluence integration by @aantn in #1617
• Add instructions for running cloud-provider MCPs via cli by @aantn in #1641
• Add Coralogix MCP toolset configurations for test cases by @aantn in #1652
• Add total tokens column to GitHub Actions eval report by @aantn in #1608
• Add comprehensive Python SDK reference documentation by @aantn in #1644
• add resourcequotas role to holmes default service account by @arikalon1 in #1653
• Add webhook configuration permissions to HolmesGPT service account by @aantn in #1655
• Add Opus 4.6 and Sonnet 4.6 to benchmark model defaults by @arikalon1 in #1624
• AWS_BEARER_TOKEN_BEDROCK for bedrock model by @arikalon1 in #1656
• Fix test comment placement in HTTP API documentation by @aantn in #1661
• docs: Expand OpenRouter configuration with Helm chart examples by @aantn in #1663
• ROB-3151 - Add interactive TUI for toolset configuration by @naomi-robusta in #1639
• Add appProtocol field to Holmes service port configuration by @aantn in #1666
• Support recursive parsing of nested schemas in MCP tool parameters by @aantn in #1668
• Prevent tool output duplication in LLM messages by @aantn in #1673
• Add direct Prometheus endpoint option for Grafana Cloud by @aantn in #1672
• Add documentation entries for new builtin toolsets by @aantn in #1679
• Clear Sentry DSN from Helm values configuration by @aantn in #1686
• Update project description and contributors in README by @aritraghosh in #1691
• HTTP Header propogation by @aantn in #1675
• Remove troubleshooting documentation from operator nav by @aantn in #1683

Full Changelog: 0.20.0...0.21.0

0.20.0

Major upgrades to HolmesGPT 🚀

Breaking Changes

• Remove additional instructions support by @naomi-robusta in #1477
• Remove unused Coralogix label settings and flatten functions by @aantn in #1473

What's Changed

New Toolsets

• Generic HTTP connector toolset by @arikalon1 in #1531
• Kubevela built-in toolsets by @mailprak in #1512
• Inspektor Gadget toolset by @mqasimsarfraz in #1506
• SQL Database Toolsets by @arikalon1 in #1568
  • PostgreSQL
  • Clickhouse
  • MariaDB
  • MySQL
  • SQLServer
  • SQLite
• MCP Servers Integrations
  • Grafana MCP by @Avi-Robusta in #1571
  • Sentry MCP by @arikalon1 in #1626
  • Prefect MCP
  • Kubernetes Remediation MCP - Run remediation actions with HolmesGPT
  • Confluence MCP

More Enhancements

• ROB-3040 gcp mcp workload identity by @RoiGlinik in #1451
• Add filesystem storage for large tool results by @Sheeproid in #1529
• Add OpenRouter to AI providers documentation by @aantn in #1554
• fix(mcp): Support nullable types in ToolParameter validation by @Mr-Neutr0n in #1501
• Add multi-cluster support and external changes handling to Robusta tools by @aantn in #1478

And much more

New Contributors

• @naomi-robusta made their first contribution in #1423
• @mailprak made their first contribution in #1512
• @mqasimsarfraz made their first contribution in #1506
• @Mr-Neutr0n made their first contribution in #1501
• @svrnm made their first contribution in #1577

Full Changelog: 0.19.0...0.20.0

0.20.4-alpha

What's Changed

• ROB-3040 GCP mcp docs make it clear what are vars and what is string in the commands by @RoiGlinik in #1533
• Add behavior_controls to /api/chat for controlling prompt components by @Sheeproid in #1479
• Rename Kafka config fields for consistency and add backward compatibility by @naomi-robusta in #1491
• Rename Datadog config fields for consistency and clarity by @naomi-robusta in #1493
• Rename headers to additional_headers in Prometheus toolset by @naomi-robusta in #1494
• Rename Elasticsearch config fields for consistency by @naomi-robusta in #1497
• Simplify MCP servers documentation and improve clarity by @aantn in #1480
• Service Now - Adds health check table configuration by @Sheeproid in #1537
• Add HTTP toolset for generic API requests to whitelisted endpoints by @arikalon1 in #1531
• Add config file change detection to auto-refresh toolsets by @naomi-robusta in #1535
• [ROB-3098] mcp servers update + cve patches by @Avi-Robusta in #1541
• Prevent agents from running pre-commit/ruff/mypy unless explicitly asked by @Sheeproid in #1545
• Add filesystem storage for large tool results by @Sheeproid in #1529
• Add OpenRouter to AI providers documentation by @aantn in #1554
• Add URL-based tab selection for MkDocs Material tabs by @aantn in #1552
• remove litellm logging handlers to remove duplications from the logs … by @arikalon1 in #1556
• Enhance toolset integration guidance with API toolset option by @aantn in #1560
• feat: Add kubevela to built-in toolsets by @mailprak in #1512
• Improve error handling and logging for LLM calls and streaming by @aantn in #1559
• Add multimodal content support to message tag parsing by @aantn in #1562
• Unwrap ExceptionGroup to surface root MCP errors by @aantn in #1561
• Prepend OOM hint before output instead of appending after by @aantn in #1564
• toolset: Add Inspektor Gadget by @mqasimsarfraz in #1506
• Add bash whitelist to cli by @aantn in #1522
• Fix ToolsetDBModel crash from model_dump unpacking all Toolset fields by @Sheeproid in #1546
• fix(mcp): Support nullable types in ToolParameter validation by @Mr-Neutr0n in #1501
• [ROB-3261] Add Grafana MCP server support to eval framework and documentation by @Avi-Robusta in #1571
• [ROB-0000] docs mcp grafana fix by @Avi-Robusta in #1572
• [ROB-3208] updating teams bot docs by @Avi-Robusta in #1540
• [ROB-0000] fixing grafana docs by @Avi-Robusta in #1573
• Increase default max_points to 500 and allow LLM override up to 1000 by @aantn in #1566
• Allow loops, conditionals, and subshells with user approval by @aantn in #1511
• remove deprecated options from docs by @aantn in #1521
• Fix documentation inconsistencies and mistakes for new users by @aantn in #1576
• Improve Prometheus rules endpoint documentation by @aantn in #1580
• Add multi-cluster support and external changes handling to Robusta tools by @aantn in #1478
• Automatically test curl commands user-facing docs by @aantn in #1412
• Remove duplicate test fixtures and consolidate context_window tags by @aantn in #1585
• docs: add Azure AD authentication methods for Azure OpenAI by @aantn in #1503
• Reorganize toolsets documentation with categories and update Azure MCP by @aantn in #1587
• Docs: add github provider by @aantn in #1370
• Add instructions for setup to CONTRIBUTING.md by @svrnm in #1577
• Skips evals on docs changes by @aantn in #1594
• Update HolmesGPT documentation link to new URL by @aantn in #1598
• Inject meaningful version into dev Docker builds via git describe by @aantn in #1595
• Add PR label support for eval selection and rename markers to tags by @aantn in #1599
• Skip build-and-test on docs-only changes, except HTTP API docs by @aantn in #1600
• Only run CLI Performance Benchmark on source changes by @aantn in #1601
• Increase liveness probe failure threshold for Holmes deployment by @aantn in #1602
• Reduce Supabase timeout from 1 hour to 1 minute by @aantn in #1581
• Clarify namespace configuration for AWS MCP server setup by @aantn in #1578
• Fix PRs blocked by skipped build-and-test check by @aantn in #1611
• Add toolsets_matrix support for comparing toolset configs on same eval scenarios by @aantn in #1609
• Add checks logic by @moshemorad in #1377
• Add metadata to HolmesStatus by @moshemorad in #1591
• Add versioned documentation support with mike by @aantn in #1536
• Enable parallel test execution with pytest-xdist by @aantn in #1621
• Suppress stderr output from ulimit commands by @arikalon1 in #1623
• Support multiple clusters and external filtering in issues metadata by @arikalon1 in #1625
• ROB-3335 Add interrupt support for LLM calls via Escape key by @naomi-robusta in #1583
• Add SQL database toolset for read-only database queries by @arikalon1 in #1568
• Remove LLM test fixtures and update test configuration by @aantn in #1631
• Add MCP server integrations for Sentry, Prefect, Kubernetes Remediation, and Confluence by @arikalon1 in #1626

New Contributors

• @mailprak made their first contribution in #1512
• @mqasimsarfraz made their first contribution in #1506
• @Mr-Neutr0n made their first contribution in #1501
• @svrnm made their first contribution in #1577

Full Changelog: 0.20.3-alpha...0.20.4-alpha

0.20.3-alpha

What's Changed

• ROB-2543 store json schema to supabase instead of yaml by @naomi-robusta in #1423
• refresh the toolsets to the database, if a tool status is changed by @arikalon1 in #1439
• Add deepseek to benchmark and setup fixes by @Sheeproid in #1435
• Weekly Benchmark Results 2026-01-27_16-11 by @github-actions[bot] in #1438
• Update Slack integration notes to reference Robusta.dev platform by @aantn in #1434
• Add Coralogix API keys to eval regression workflow by @aantn in #1442
• Clarify GCP project/region settings are defaults, not access restrictions by @aantn in #1443
• ROB-2506 use litellm built in cache control by @RoiGlinik in #1436
• Add Model File Support to evals and benchmarks by @Sheeproid in #1445
• Weekly Benchmark Results 2026-01-29_09-48 by @github-actions[bot] in #1449
• bump litellm and open ai versions to latest stable by @Sheeproid in #1450
• ROB-3040 gcp mcp workload identity by @RoiGlinik in #1451
• ROB-1322 splunk mcp docs by @RoiGlinik in #1419
• Add Datadog API keys to eval regression workflow by @aantn in #1455
• Remove deprecated workload health by @aantn in #1380
• Support using config file in server mode by @gossion in #1430
• Refocus OpenAI-compatible docs on LiteLLM Proxy gateway pattern by @aantn in #1458
• Add eval test for Elasticsearch trace query with large fields by @aantn in #1461
• Improve elasticsearch_search source parameter documentation by @aantn in #1378
• Test cli performance by @aantn in #1415
• Add additional Datadog whitelist endpoints by @aantn in #1454
• Allow running Prometheus evals in CI/CD by @aantn in #1457
• Fix Datadog API endpoint patterns and add missing descriptions by @aantn in #1467
• Use hardcoded kubectl version to improve setup reliability by @aantn in #1468
• ROB-2543 refactor: standardize toolset config field names with backward compat… by @naomi-robusta in #1425
• Refactor cli and chat prompt to be build in the same function + allow controlling prompt components by @Sheeproid in #1452
• Add ENV_CONFIGS support for evaluation tests & fix kubectl tabular query filtering by @Sheeproid in #1471
• Improve various MCP docs by @aantn in #1304
• Remove old alert-to-runbook YAML mapping system by @aantn in #1404
• Improve prometheus toolset (alert listing tool) by @aantn in #1352
• Remove additional instructions support by @naomi-robusta in #1477
• Add release version to Sentry initialization by @moshemorad in #1499
• Rename New Relic config fields for consistency by @naomi-robusta in #1492
• Rename ServiceNow instance_url to api_url with backward compatibility by @naomi-robusta in #1496
• fix cves for several packages by @mainred in #1481
• Remove unused Coralogix label settings and flatten functions by @aantn in #1473
• Migrate Datadog evals from EU to US5 site and remove stale mocked test by @aantn in #1507
• minor runbook eval improvements by @aantn in #1520
• Add filtering and limit parameters to list active metrics tool by @aantn in #1466
• Add icon_url support for MCP servers by @aantn in #1463
• Add Git workflow guidelines to contributor documentation by @naomi-robusta in #1525
• feat: publish prebuilt CLI binaries with standard naming convention by @aantn in #1422
• Fix Windows build by excluding jiter 0.13.0 by @moshemorad in #1528
• Big yaml eval by @Sheeproid in #1527
• Rename RabbitMQ config fields for consistency and add backward compatibility by @naomi-robusta in #1495
• Rename Grafana config fields for consistency and add backward compatibility by @naomi-robusta in #1498

New Contributors

• @naomi-robusta made their first contribution in #1423

Full Changelog: 0.19.1...0.20.3-alpha

0.20.2-alpha

What's Changed

• ROB-2543 store json schema to supabase instead of yaml by @naomi-robusta in #1423
• refresh the toolsets to the database, if a tool status is changed by @arikalon1 in #1439
• Add deepseek to benchmark and setup fixes by @Sheeproid in #1435
• Weekly Benchmark Results 2026-01-27_16-11 by @github-actions[bot] in #1438
• Update Slack integration notes to reference Robusta.dev platform by @aantn in #1434
• Add Coralogix API keys to eval regression workflow by @aantn in #1442
• Clarify GCP project/region settings are defaults, not access restrictions by @aantn in #1443
• ROB-2506 use litellm built in cache control by @RoiGlinik in #1436
• Add Model File Support to evals and benchmarks by @Sheeproid in #1445
• Weekly Benchmark Results 2026-01-29_09-48 by @github-actions[bot] in #1449
• bump litellm and open ai versions to latest stable by @Sheeproid in #1450
• ROB-3040 gcp mcp workload identity by @RoiGlinik in #1451
• ROB-1322 splunk mcp docs by @RoiGlinik in #1419
• Add Datadog API keys to eval regression workflow by @aantn in #1455
• Remove deprecated workload health by @aantn in #1380
• Support using config file in server mode by @gossion in #1430
• Refocus OpenAI-compatible docs on LiteLLM Proxy gateway pattern by @aantn in #1458
• Add eval test for Elasticsearch trace query with large fields by @aantn in #1461
• Improve elasticsearch_search source parameter documentation by @aantn in #1378
• Test cli performance by @aantn in #1415
• Add additional Datadog whitelist endpoints by @aantn in #1454
• Allow running Prometheus evals in CI/CD by @aantn in #1457
• Fix Datadog API endpoint patterns and add missing descriptions by @aantn in #1467
• Use hardcoded kubectl version to improve setup reliability by @aantn in #1468
• ROB-2543 refactor: standardize toolset config field names with backward compat… by @naomi-robusta in #1425
• Refactor cli and chat prompt to be build in the same function + allow controlling prompt components by @Sheeproid in #1452
• Add ENV_CONFIGS support for evaluation tests & fix kubectl tabular query filtering by @Sheeproid in #1471
• Improve various MCP docs by @aantn in #1304
• Remove old alert-to-runbook YAML mapping system by @aantn in #1404
• Improve prometheus toolset (alert listing tool) by @aantn in #1352
• Remove additional instructions support by @naomi-robusta in #1477
• Add release version to Sentry initialization by @moshemorad in #1499
• Rename New Relic config fields for consistency by @naomi-robusta in #1492
• Rename ServiceNow instance_url to api_url with backward compatibility by @naomi-robusta in #1496
• fix cves for several packages by @mainred in #1481
• Remove unused Coralogix label settings and flatten functions by @aantn in #1473
• Migrate Datadog evals from EU to US5 site and remove stale mocked test by @aantn in #1507
• minor runbook eval improvements by @aantn in #1520
• Add filtering and limit parameters to list active metrics tool by @aantn in #1466
• Add icon_url support for MCP servers by @aantn in #1463
• Add Git workflow guidelines to contributor documentation by @naomi-robusta in #1525
• feat: publish prebuilt CLI binaries with standard naming convention by @aantn in #1422

New Contributors

• @naomi-robusta made their first contribution in #1423

Full Changelog: 0.19.0...0.20.2-alpha

0.19.1

What's Changed

• Upgrade starlette from 0.46.0 to 0.49.1 to fix
  https://github.com/advisories/GHSA-7f5h-v6xp-fcq8
• Upgrade urllib3 from 1.26.20 to 2.6.3 to address
  https://github.com/advisories/GHSA-2xpw-w6gg-jr37,
  https://github.com/advisories/GHSA-gm62-xv2j-4w53 and
  https://github.com/advisories/GHSA-38jv-5279-wg99
• Upgrade prometrix to 0.2.9 to allow bump urllib3. prometrix 0.2.5
  relies on urllib3 (>=1.26.20,<2.0.0)
• Bump fastapi because fastapi (0.116.0) depends on starlette
  (>=0.40.0,<0.47.0)

Full Changelog: 0.19.0...0.19.1

0.19.0

New Features

• Enhanced and refined the Bash toolset.
• Added support for daily and weekly scheduled prompt investigations in SAS, available starting with this release.

New mcp servers

• GitHub MCP

  • Introduced by replacing the legacy GitHub toolset with an MCP-based implementation #1401

• Confluence MCP

  • Used for Confluence evaluations and workflows #1395 #1407

• Splunk MCP

  • Added explicit Splunk MCP server support PR: #1399

Breaking Changes

The git toolset has been deprecated in favor of the Git MCP server see Docs here

What's Changed

• Add secrets to GitHub Actions for confluence evals by @aantn in #1382
• [ROB-3060] removed api key from pytests by @Avi-Robusta in #1389
• Add gemini api to evals and fixed test in merging them by @Sheeproid in #1391
• Add OpenRouter docs by @aantn in #1386
• add fast benchmark result with gpt-5.2 by @Sheeproid in #1394
• Confluence evals by @aantn in #1395
• ROB-1322 splunk mcp by @RoiGlinik in #1399
• doc improvments by @aantn in #1371
• Remove old GitHub toolset in favor of MCP-based implementation by @aantn in #1401
• Add support for image input by @aantn in #1348
• Add images field to chat API documentation by @aantn in #1403
• Weekly Benchmark Results 2026-01-22_07-43 by @github-actions[bot] in #1406
• Add CONFLUENCE_USER_NAME env var to eval-regression workflow by @aantn in #1407
• Remove default runbooks from catalog by @Sheeproid in #1409
• Better oomkill detection by @aantn in #1408
• using more models for scheduled eval benchmark run by @Sheeproid in #1417
• remove disabling non existing tools aws/rds and aws/security by @Sheeproid in #1418
• Redesigned bash toolset that have a prefix based approvals by @Sheeproid in #1360
• feat: report json schema of toolset config to db by @aantn in #1327
• fix typo by @aantn in #1424
• feat: support passing token to mcp in header by @gossion in #1397
• [ROB-2990] Scheduled prompts support by @Avi-Robusta in #1420

New Contributors

• @github-actions[bot] made their first contribution in #1406
• @gossion made their first contribution in #1397

Full Changelog: 0.18.5...0.19.0