You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I think push vs pull has been mentioned a long time ago, on numerous occasions, and in numerous sources already, but after briefly reviewing open (and closed) issues I haven't found anything of a decisive "yes" / "no" / "maybe" / "send a patch".
The high-level proposal is to expose an endpoint, where clients could push metrics directly, ideally via HTTP POST. Basically integrate (parts of) pushgateway into "core" Prometheus, and make push-based metric collection a first-class citizen.
Clients that only expose metrics via the scrape/GET endpoint, could be bridged by a local scraper that later pushes the metrics via POST.
The text was updated successfully, but these errors were encountered:
it might be worth doing some search as I suspect this topic has already been discussed.
Don't forget to describe your use case and why it can't be solved with the pull model or via the push gateway.
I think push vs pull has been mentioned a long time ago, on numerous occasions, and in numerous sources already, but after briefly reviewing open (and closed) issues I haven't found anything of a decisive "yes" / "no" / "maybe" / "send a patch".
From what I gathered, pushgateway, due to its design, is only ever recommended for certain, specialized use cases.
The high-level proposal is to expose an endpoint, where clients could push metrics directly, ideally via HTTP POST. Basically integrate (parts of) pushgateway into "core" Prometheus, and make push-based metric collection a first-class citizen.
Clients that only expose metrics via the scrape/GET endpoint, could be bridged by a local scraper that later pushes the metrics via POST.
The text was updated successfully, but these errors were encountered: