Fix #917 — base FinOps memory recommendation on 7-day P95#918
Merged
erikdarlingdata merged 1 commit intodevfrom May 2, 2026
Merged
Fix #917 — base FinOps memory recommendation on 7-day P95#918erikdarlingdata merged 1 commit intodevfrom
erikdarlingdata merged 1 commit intodevfrom
Conversation
…shot Both the Memory right-sizing check (#3) and the VM right-sizing memory prescription (#12) read instantaneous values, which produced misleading "buffer pool uses 0%" findings on servers that were either freshly restarted or under genuine memory pressure where plan cache / workspace memory dominate. - Check #3 now reads P95 of collect.memory_stats.total_memory_mb over 7 days instead of TOP(1) buffer_pool_mb at the latest collection. - Check #12 replaces the live perfmon "Database Cache Memory (KB)" read (data-cache slice only, instantaneous) with the same 7-day P95 of total_memory_mb. CPU side already uses 7-day P95. - Both checks require >= 500 samples (~1 day at 1/min) before firing. - Recommendation text now says "P95 SQL memory" rather than "buffer pool" to reflect what is actually being measured. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This was referenced May 2, 2026
pull Bot
pushed a commit
to ehtick/PerformanceMonitor
that referenced
this pull request
May 5, 2026
…ests PRs erikdarlingdata#918 and erikdarlingdata#920 added a sample-count guard (>= 500) to the new 7-day P95 memory recommendation logic. 500 was overly conservative and broke the Lite FinOps test OverProvisionedEnterprise_MemoryRightSizingFires, which seeds a single memory_stats row. The real protection added by those PRs was switching from TOP 1 to P95; the sample minimum is just a sanity check against degenerate single-point inputs. ~16 samples is enough to compute a meaningful P95 and matches the shape of SeedCpuUtilizationAsync's 16-row fixture, so tests can fire the recommendation without artificial inflation. - Lower threshold from 500 to 16 in both Dashboard and Lite (checks #3 and erikdarlingdata#12) so the value reflects the actual ask: "more than one reading" rather than "8+ hours of data." - Update Lite's SeedMemoryStatsAsync to insert 16 rows across the test period (matching SeedCpuUtilizationAsync's pattern). This makes OverProvisionedEnterprise_MemoryRightSizingFires pass again and keeps CleanServer_NoDuckDbRecommendations green (still no rows seeded for the clean scenario → P95 returns NULL → no recommendation). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
collect.memory_stats.total_memory_mb— the same signal the Utilization tab reads — and requires ~1 day of samples before firing.Fixes #917
Test plan
🤖 Generated with Claude Code