Baoyuantop commented on issue #11948:
URL: https://github.com/apache/apisix/issues/11948#issuecomment-4043851939
Here are some ways to alleviate this problem:
1. Configuration metrics expiration time
```
plugin_attr:
prometheus:
metrics:
http_status:
expire: 600 # 10 mins
http_latency:
expire: 600
bandwidth:
expire: 600
upstream_status:
expire: 300 # 5 mins
```
2. Reduce the use of extra_labels, or remove unnecessary high-cardinality
labels. Not configuring extra_labels in Prometheus's plugin_attr can
significantly reduce the number of unique indicator keys.
3. Increase the capacity of a shared dictionary
```
nginx_config:
http:
lua_shared_dict:
prometheus-metrics: 100m
```
4. Upgrade to version 3.14.0+
The architecture improvements obtained from PR #12383 (decoupling
computation and output) reduce repetitive operations and memory pressure
between worker processes.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
