CPU and Memory full on Harbor database server

[rhegde2]

Hi Team,

We are using Harbor Version v2.9.1

When we trigger the image scan, we are experiancing sudden increase in number of database connections which inturn uses most of the CPU and Memory. It turns the DB server to hung state.

We tried increasing the CPU and RAM, and no luck.

There is no error on Postgres logs, however it was showing lot of "UPDATE waiting" message during the issue.

Below updates were running during the activity

---

UPDATE TAK

UPDATE "schedule" SET "revision"

UPDATE "repository" SET "pull_count" = "pull_count" + $1, "update_time" = $2 WHERE "repository_id" IN ( SELECT T0."repository_id" FROM "repository" T0 WHERE T0."repository_id" = $3;

---

Kindly suggest.

[chlins]

Could you provide more detailed data such as screenshot of the monitoring, harbor logs？

[stonezdj]

Suggest upgrading to the latest Harbor 2.12.0 because there are lots of performance improvements after v2.9.0. Is there any repository which has too many artifact or tags (>1000) ?

[chlins]

@rhegde2 Did you change any environment variables manually? the pull count updation should be async by default since v2.5, related PR: #16107