Integration of pgwatch into a large data center #692
-
|
Hello, I am conducting a PoC for pgwatch in a large data center. I am using the Docker container version of pgwatch. It works smoothly when I add around 600-700 databases. However, when I try to monitor 2000 databases with a single pgwatch instance, I experience performance issues and receive metrics with delays. When I monitor server resources, CPU and RAM usage appear to be low. Is there a way to increase the number of pgwatch worker processes? Or would you recommend horizontal scaling for monitoring at this scale? I need your recommendations. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
Would you please provide log output for investigations? Each monitored source receives it's own worker. There are no limitations here. |
Beta Was this translation helpful? Give feedback.
-
|
Hello, the problem was not with the pgwatch3 worker. There was a disk latency problem because I added too many databases. I solved the problem by distributing the database inventory to different pgwatch3 servers. Thanks |
Beta Was this translation helpful? Give feedback.
Hello, the problem was not with the pgwatch3 worker. There was a disk latency problem because I added too many databases. I solved the problem by distributing the database inventory to different pgwatch3 servers. Thanks