|
1 | | -### Q4 2025 Objectives |
| 1 | +### Q1 2026 Objectives |
2 | 2 |
|
3 | | -#### Go all in on AI (Driver: <TeamMember name="David Newell" />) |
4 | | -* Implement issue autofixing |
5 | | - * In-app flow where an issue can be assigned to Max |
6 | | - * Build terminal based experience within the CLI |
7 | | - * Add issue selection to MCP (e.g. by file, most impactful) |
8 | | - * Stretch: Add test writing agent to prevent regressions |
9 | | -* Add exception capture auto instrumentation to the wizard |
10 | | -* Integrations to capture exceptions for popular LLM / agent tools (e.g. Langchain) |
| 3 | +#### **Alerting** - Spikes & Digest (Driver: <TeamMember name="Aleksander Błaszkiewicz" />) |
| 4 | +* Define how we calculate a spike and what are the controls |
| 5 | +* Add 'Issue Spiked' trigger in alerting |
| 6 | +* Add ability to snooze issues from alerting |
| 7 | +* Create weekly digest (warnings, resolved, recommendations, priorities...) |
| 8 | +* Stretch: correlate with event traffic to understand the context of a “spike” |
11 | 9 |
|
12 | | -#### Ship mobile error tracking (Driver: <TeamMember name="Manoel Aranda Neto" />) |
13 | | -* Complete 'straightforward' section of the [megaissue](https://github.com/PostHog/posthog/issues/38845) |
14 | | -* Support suppression rules |
15 | | -* Add suppression of minified exceptions before cymbal processing |
16 | | -* Build dashboard template including crash free sessions / users metric |
| 10 | +#### **Engagement** - Search & Speed (Driver: <TeamMember name="Aleksander Błaszkiewicz" />) |
| 11 | +* Move issues to CH to avoid dual DB query |
| 12 | +* Stretch: Improve loading speed on search results |
17 | 13 |
|
18 | | -#### Research deep dive on customers (Driver: <TeamMember name="Cory Slater" />) |
19 | | -* Interview existing, potential, and recently churned customers to understand: |
20 | | - * What makes people switch from competitors and what pulls them back? |
21 | | - * What value would convince customers to remove billing limits? |
22 | | - * How do AI-first companies handle errors differently? |
23 | | - * What context would let customers skip steps in their workflow? |
24 | | - * What drives month to month swings in spend vs contraction? |
25 | | -* Every engineer to attend at least 2 customer interviews |
| 14 | +#### **Grouping** - Fingerprint & Sourcemaps (Driver: <TeamMember name="Hugues Pouillot" />) |
| 15 | +* Iterate on fingerprint algorithm and compare with existing approach using rand index |
| 16 | +* Stretch: Add bundler plugins to make sourcemap uploads easier |
| 17 | +* Stretch: Explore embedding based grouping |
26 | 18 |
|
27 | | -#### Increase trust in product (Driver: <TeamMember name="Hugues Pouillot" />) |
28 | | -* Improve issue grouping |
29 | | - * Introduce a measure of grouping accuracy to base changes off |
30 | | - * Add fingerprint versioning to avoid duplicate issues when iterating |
31 | | - * Fix obvious mis-groupings based on minor stack trace variances |
32 | | - * Ship embeddings to support merge suggestions |
33 | | -* Decide with Team ClickHouse on approach to avoid dual DB querying |
34 | | -* Reduce the number of stacktrace-less issues |
35 | | -* Give better context when a recording is not found / source maps missing |
36 | | -* Rethink UI / UX of issue page |
| 19 | +#### **Collection** - Cymbal refactoring & Mobile support (Driver: <TeamMember name="Hugues Pouillot" />) |
| 20 | +* Expose frame resolution and issue processing as endpoints |
| 21 | +* Coordinate with Ingestion team to move cymbal processing into plugin server |
| 22 | +* Make CLI compatible with stable chunk ids to: |
| 23 | + * be compatible with debug id specification and reuse current tooling. |
| 24 | + * avoid rebuilding application |
| 25 | + * speed up ingestion by reducing symbol set downloads |
| 26 | + * reduce symbol set uploads |
| 27 | +* Add debug symbol upload command to CLI |
| 28 | +* Support debug symbols with cymbal |
37 | 29 |
|
38 | | -#### Improving context on issues (Driver: <TeamMember name="Aleksander Błaszkiewicz" />) |
39 | | -* Add breakdowns to issues (by release / person_id) |
40 | | -* Show first affected release & relevant commit |
41 | | - * Stretch: Auto assign to commit author |
42 | | -* Level up Node SDK (add contexts & tagging, $session_id propagation) |
43 | | -* Capture code variables in Python |
44 | | -* Stretch: Integrate logs into session timeline |
45 | | - |
46 | | -#### Stretch: Billing, spikes and alerts (Driver: <TeamMember name="Oliver Browne" />) |
47 | | -(Note: this goal largely depends on the capacity Olly has outside of mobile error tracking, which takes precedence) |
48 | | -* **Email** alerts for total exceptions ingested volume spikes |
49 | | -* Add billing limits to onboarding |
50 | | -* Fixed issue sampling (manually configured, drop X% of exceptions for this issue) |
51 | | - * Stretch: Adaptive sampling (spread exception capture over billing period) |
52 | | -* Threshold based alerting per-issue, based on product analytics insight alerts |
53 | | - * Support "X occurrences per Y duration" |
54 | | - * Add to alerting step during onboarding |
55 | | - * Stretch: Support "X affected users per Y duration" |
56 | | -* Issue time-based snoozing (drop all exceptions for X hours/days, re-open the issue if seen after that period) |
| 30 | +#### **Monitoring** - Acquisition & KPIs (Driver: <TeamMember name="Cory Slater" />) |
| 31 | +* Define KPIs for tracking collection, grouping, alerting, engagement |
| 32 | +* Deep dive into acquisition metrics |
| 33 | +* Case study around successful migration |
0 commit comments