To avoid dup alert, any guide doc to help tweak --cluster.gossip-interval and --cluster.pushpull-interval

[jicahoo]

Hi ,

Thanks in advance for any feedback for my question. 😄 I am newbie to distributed system. So my question may be stupid.
Now our AM config group_interval is 1m. However, it will generate dup alert in our AlertManager cluster. In order to avoid dup alert, I increased group_interval to 5m. However, my team's architect disagree with that "5m is too long. Smaller group_interval can make SRE to check the issue ASAP".

So I'm wondering if there any detailed guide to tweak the cluster options and group_intervals to avoid dup alert?

I searched on the Internet but not found a official guide. Now I am reading the code of alertmanager to figure out the mechanism of dedup alert 😄

web pages I found about alertmanager alert dedup:
-- https://groups.google.com/g/prometheus-users/c/8bnd_-BJx68
-- #2222
-- #1550
-- #1982
-- https://github.com/prometheus/alertmanager#high-availability

Or anyone can help answer what's difference between gossip-inerval and pushpll-interval? Both of them will impact the replication of notification log?