Issues

Managed role provisioning re-sends every role's plaintext password each start; a raising DO block puts all three in the store log, which viewer can read

#3910

· erikdarlingdata opened

on Sep 22, 2026

Managed store writes maintenance_work_mem = 2048MB, one kB over PostgreSQL 17's Windows maximum: a 17.x store on a 40 GB+ host cannot restart

bug

#3909

· erikdarlingdata opened

on Sep 22, 2026

TimescaleDB bump: the store upgrade cannot move a field store's extension (same-major probe loads the missing old library; pg_upgrade bridge cannot reach a newer version)

bug

#3908

· erikdarlingdata opened

on Sep 22, 2026

Security: bundled PostgreSQL 18.4 / TimescaleDB 2.28.1 and the self-contained .NET 10.0.11 runtime carry published CVEs

bug

dependencies

#3906

· erikdarlingdata opened

on Sep 22, 2026

Web viewer server page: the two daily-summary reads are ~26 of the Overview tab's ~28 s on a busy production primary, one at the viewer role's 15 s timeout — raw CTEs over the whole window plus a per-call rollup-coverage probe

#3905

· erikdarlingdata opened

on Sep 22, 2026

get_store_metrics returns ~1.9 MB by default (every store object × ~30 days) — ~500k tokens, larger than any agent context; even days_back=1 is 346 KB

#3903

· erikdarlingdata opened

on Sep 22, 2026

analyze_server takes 15–21 s on the production SQL store that ingests Query Store, dominated by PlanRegressionSql's per-call 14-day dedup — not the latest-value reads #3896 measured

#3902

· erikdarlingdata opened

on Sep 22, 2026

PostgreSQL-target analysis re-reads 30 days of statement stats once per keyed statement: cold analyze_server takes 5–31 s on the production PG store, ~70 % of it in the keyed baselines

#3901

· erikdarlingdata opened

on Sep 22, 2026

The MCP surface costs ~100k tokens of context before the first call: 157 tools, 336 KB of tools/list, 86 KB of instructions

#3898

· erikdarlingdata opened

on Sep 22, 2026

Trend tools return every raw point: get_file_io_trend sends 12,451 points / 1.4 MB for one server at defaults, more than an LLM client's context

#3897

· erikdarlingdata opened

on Sep 22, 2026

Analysis 'latest value' reads scan a server's entire retained history: half of every analyze_server call, and the database-size fact counts dropped databases (all three SKUs)

#3896

· erikdarlingdata opened

on Sep 22, 2026

Fleet overview reads every retained row of four tables on every call, and the web viewer calls it twice a minute per open tab

#3895

· erikdarlingdata opened

on Sep 22, 2026

Navigation Menu

All issues

Managed role provisioning re-sends every role's plaintext password each start; a raising DO block puts all three in the store log, which viewer can read

Managed store writes maintenance_work_mem = 2048MB, one kB over PostgreSQL 17's Windows maximum: a 17.x store on a 40 GB+ host cannot restart

TimescaleDB bump: the store upgrade cannot move a field store's extension (same-major probe loads the missing old library; pg_upgrade bridge cannot reach a newer version)

Security: bundled PostgreSQL 18.4 / TimescaleDB 2.28.1 and the self-contained .NET 10.0.11 runtime carry published CVEs

Web viewer server page: the two daily-summary reads are ~26 of the Overview tab's ~28 s on a busy production primary, one at the viewer role's 15 s timeout — raw CTEs over the whole window plus a per-call rollup-coverage probe

get_store_metrics returns ~1.9 MB by default (every store object × ~30 days) — ~500k tokens, larger than any agent context; even days_back=1 is 346 KB

analyze_server takes 15–21 s on the production SQL store that ingests Query Store, dominated by PlanRegressionSql's per-call 14-day dedup — not the latest-value reads #3896 measured

PostgreSQL-target analysis re-reads 30 days of statement stats once per keyed statement: cold analyze_server takes 5–31 s on the production PG store, ~70 % of it in the keyed baselines

The MCP surface costs ~100k tokens of context before the first call: 157 tools, 336 KB of tools/list, 86 KB of instructions

Trend tools return every raw point: get_file_io_trend sends 12,451 points / 1.4 MB for one server at defaults, more than an LLM client's context

Analysis 'latest value' reads scan a server's entire retained history: half of every analyze_server call, and the database-size fact counts dropped databases (all three SKUs)

Fleet overview reads every retained row of four tables on every call, and the web viewer calls it twice a minute per open tab

Search results