Infisical is the open-source platform for securing secrets, certificates, access, and AI Agents.
TypeScript acmeagentsai-agentscertificate-managementdevsecopsenvironment-variableskubernetesopen-sourcepampkipostgresprivate-caprivileged-access-managementsecret-managementsecret-managersecretssecrets-managementsecuritysecurity-toolsvault
refactor: update cron job frequency for SecretReadActivityPrune to daily
Dimitris Staratzis8674938dc8b5
diff --git a/backend/CLAUDE.md b/backend/CLAUDE.mdindex 5eb2daf989..9c4e254a0e 100644--- a/backend/CLAUDE.md+++ b/backend/CLAUDE.md@@ -1097,7 +1097,7 @@ and the events cannot say which folders a recursive read returned. Instead it kenon-blocking lock), by the `SecretReadActivityFlush` cron every minute into an upsert.- **Partitioned by month, no foreign keys, no index on `lastReadAt`.** The table is range-partitioned on`readMonth` (the UTC month of `lastReadAt`, part of the primary key), so a reader has one row per month and- `findLastReads` folds them into the latest. The hourly `SecretReadActivityPrune` cron creates the current and+ `findLastReads` folds them into the latest. The daily `SecretReadActivityPrune` cron creates the current andnext three months' partitions and drops a month once all of it is older than the 30-day window, since nothingreads before it. Never prune with `DELETE`: at this table's size that is a full scan plus vacuum. Partition DDLtakes an exclusive lock on the whole table, so it runs under a short `lock_timeout` and retries next hour.diff --git a/backend/src/ee/services/secret-read-activity/secret-read-activity-service.ts b/backend/src/ee/services/secret-read-activity/secret-read-activity-service.tsindex e5c4170a1d..91b37228d5 100644--- a/backend/src/ee/services/secret-read-activity/secret-read-activity-service.ts+++ b/backend/src/ee/services/secret-read-activity/secret-read-activity-service.ts@@ -410,12 +410,12 @@ export const secretReadActivityServiceFactory = ({enabled: !appCfg.isSecondaryInstance,handler: $consumeReadStream});- // Hourly although a month's partition is needed once, so that an instance back from a long outage gets- // partitions for its reads within the hour.+ // Daily although a month's partition is needed once: partitions exist months ahead, so a missed run costs+ // nothing, and an instance back from a longer outage gets them within a day while its reads wait in the stream.cronJob.register({name: CronJobName.SecretReadActivityPrune,- pattern: "15 * * * *",- runHashTtlS: 24 * 60 * 60,+ pattern: "15 1 * * *",+ runHashTtlS: 3 * 24 * 60 * 60,enabled: !appCfg.isSecondaryInstance,handler: async () => {const created = await secretReadActivityDAL.createMonthlyPartitions(