[ROB-2347]improve prompt cacheing

[Avi-Robusta]

No description provided.

[coderabbitai]

Walkthrough

The PR relocates the _current_date_time.jinja2 template includes from the top to the bottom of six prompt template files, adjusting the timing of when current date/time context is injected during template rendering. One file also includes a minor clarifying sentence.

Changes

Cohort / File(s)	Change summary
Generic ask templates holmes/plugins/prompts/generic_ask.jinja2, generic_ask_conversation.jinja2, generic_ask_for_issue_conversation.jinja2	Relocated _current_date_time.jinja2 include from top to end of template (after conditional blocks or main content).
Kubernetes workload templates holmes/plugins/prompts/kubernetes_workload_ask.jinja2, kubernetes_workload_chat.jinja2	Relocated _current_date_time.jinja2 include from top to end of template. kubernetes_workload_chat.jinja2 also adds a sentence clarifying that not all questions may relate to prior investigation.
Investigation template holmes/plugins/prompts/generic_investigation.jinja2	Relocated _current_date_time.jinja2 include from top to end of template (after investigation_output_format.jinja2).

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

The changes follow a consistent, homogeneous pattern across all affected files—relocating template includes without introducing new logic or complex interactions. The repetitive nature of the edits reduces cognitive load per file, though verification across multiple files is necessary.

Possibly related PRs

• ROB-1441 global date context #552: Modifies the same prompt templates with context injection via global now/now_timestamp_seconds variables, representing an alternative approach to current date/time availability.
• chore: Add ai safety prompt to system prompt #823: Edits the same prompt template files' include directives, adding _ai_safety includes alongside the _current_date_time relocations.

Suggested reviewers

• arikalon1
• mainred

Pre-merge checks and finishing touches

❌ Failed checks (1 inconclusive)

Check name	Status	Explanation	Resolution
Description Check	❓ Inconclusive	No pull request description was provided by the author. Without any description content to evaluate, there is insufficient information to determine whether it is related to or independent from the changeset. The check instructions focus on assessing the relationship between the description content and the changes, but no content exists to perform this evaluation.	Consider adding a brief pull request description that explains the purpose of moving the _current_date_time.jinja2 template inclusions and how this change improves prompt caching performance. Even a short explanation would help reviewers understand the reasoning behind the reorganization.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title Check	✅ Passed	The title "[ROB-2347]improve prompt cacheing" is clearly related to the main changes in the pull request. The raw summary shows that across six template files, the _current_date_time.jinja2 template inclusion has been consistently moved from the top to the bottom of files, and the PR objectives confirm this change is intended to improve prompt caching. The title accurately captures the goal of the changeset, even though it doesn't explicitly describe the specific mechanism of moving the includes. This is sufficiently clear and specific for teammates scanning the history to understand the primary intent.
Docstring Coverage	✅ Passed	No functions found in the changes. Docstring coverage check skipped.

✨ Finishing touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch move-datetime

---

📜 Recent review details

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between de3a4a2 and c7dc636.

📒 Files selected for processing (6)

• holmes/plugins/prompts/generic_ask.jinja2 (1 hunks)
• holmes/plugins/prompts/generic_ask_conversation.jinja2 (1 hunks)
• holmes/plugins/prompts/generic_ask_for_issue_conversation.jinja2 (1 hunks)
• holmes/plugins/prompts/generic_investigation.jinja2 (1 hunks)
• holmes/plugins/prompts/kubernetes_workload_ask.jinja2 (1 hunks)
• holmes/plugins/prompts/kubernetes_workload_chat.jinja2 (1 hunks)

🧰 Additional context used

📓 Path-based instructions (1)

holmes/plugins/prompts/**/*.jinja2

📄 CodeRabbit inference engine (CLAUDE.md)

Prompts must be stored as Jinja2 templates under holmes/plugins/prompts/{name}.jinja2

Files:

• holmes/plugins/prompts/generic_ask_for_issue_conversation.jinja2
• holmes/plugins/prompts/generic_ask.jinja2
• holmes/plugins/prompts/kubernetes_workload_ask.jinja2
• holmes/plugins/prompts/generic_ask_conversation.jinja2
• holmes/plugins/prompts/generic_investigation.jinja2
• holmes/plugins/prompts/kubernetes_workload_chat.jinja2

🧠 Learnings (1)

📚 Learning: 2025-10-05T13:01:12.288Z

Learnt from: CR
PR: robusta-dev/holmesgpt#0
File: CLAUDE.md:0-0
Timestamp: 2025-10-05T13:01:12.288Z
Learning: Applies to holmes/plugins/prompts/**/*.jinja2 : Prompts must be stored as Jinja2 templates under holmes/plugins/prompts/{name}.jinja2

Applied to files:

• holmes/plugins/prompts/generic_ask.jinja2

⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (3)

• GitHub Check: llm_evals
• GitHub Check: Pre-commit checks
• GitHub Check: build

🔇 Additional comments (6)

holmes/plugins/prompts/generic_investigation.jinja2 (1)

43-44: LGTM - Timestamp relocation improves prompt caching.

Moving the date/time include to the end allows LLM providers to cache the static instruction prefix, since the timestamp no longer invalidates the entire prompt on each request.

holmes/plugins/prompts/kubernetes_workload_chat.jinja2 (1)

39-40: LGTM - Consistent timestamp relocation.

The timestamp include has been appropriately moved to the end, consistent with the caching optimization applied across all prompt templates in this PR.

holmes/plugins/prompts/generic_ask_for_issue_conversation.jinja2 (1)

51-52: LGTM - Timestamp moved to end for caching.

The relocation maintains functionality while enabling better prompt cache utilization.

holmes/plugins/prompts/generic_ask.jinja2 (1)

42-43: LGTM - Timestamp relocation aligns with caching strategy.

The include has been moved to the end after optional additions, maintaining the pattern of placing dynamic content at the end for optimal caching. File location complies with the project's Jinja2 template guidelines.

holmes/plugins/prompts/generic_ask_conversation.jinja2 (1)

33-34: LGTM - Consistent with PR's caching optimization.

The timestamp include has been appropriately relocated to the end of the template.

holmes/plugins/prompts/kubernetes_workload_ask.jinja2 (1)

78-79: LGTM - Final template updated for caching optimization.

The timestamp include has been moved to the end, completing the consistent refactoring across all six prompt templates for improved caching efficiency.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[github-actions]

Results of HolmesGPT evals

• ask_holmes: 34/35 test cases were successful, 0 regressions, 1 setup failures

Test suite	Test case	Status
ask	01_how_many_pods	✅
ask	02_what_is_wrong_with_pod	✅
ask	04_related_k8s_events	✅
ask	05_image_version	✅
ask	09_crashpod	✅
ask	10_image_pull_backoff	✅
ask	110_k8s_events_image_pull	✅
ask	11_init_containers	✅
ask	13a_pending_node_selector_basic	✅
ask	14_pending_resources	✅
ask	15_failed_readiness_probe	✅
ask	17_oom_kill	✅
ask	19_detect_missing_app_details	✅
ask	20_long_log_file_search	✅
ask	24_misconfigured_pvc	✅
ask	24a_misconfigured_pvc_basic	✅
ask	28_permissions_error	🚧
ask	39_failed_toolset	✅
ask	41_setup_argo	✅
ask	42_dns_issues_steps_new_tools	✅
ask	43_current_datetime_from_prompt	✅
ask	45_fetch_deployment_logs_simple	✅
ask	51_logs_summarize_errors	✅
ask	53_logs_find_term	✅
ask	54_not_truncated_when_getting_pods	✅
ask	59_label_based_counting	✅
ask	60_count_less_than	✅
ask	61_exact_match_counting	✅
ask	63_fetch_error_logs_no_errors	✅
ask	79_configmap_mount_issue	✅
ask	83_secret_not_found	✅
ask	86_configmap_like_but_secret	✅
ask	93_calling_datadog[0]	✅
ask	93_calling_datadog[1]	✅
ask	93_calling_datadog[2]	✅

Legend

• ✅ the test was successful
• :minus: the test was skipped
• ⚠️ the test failed but is known to be flaky or known to fail
• 🚧 the test had a setup failure (not a code regression)
• 🔧 the test failed due to mock data issues (not a code regression)
• 🚫 the test was throttled by API rate limits/overload
• ❌ the test failed and should be fixed before merging the PR