tsenart
Prometheus support
Hello, it is great to see metrics in prometheus, is it possible to do this? jaggr is used to plot realtime graph, so I think it should possible to convert it to prometheus format.
Any thoughts?
It would be possible to create a sub command that consumes vegeta attack results (i.e. like vegeta report or vegeta encode) that is a prometheus exporter.
Proposal
What do you think about creating a new command category called "monitor" aside report, encode, dump?
Monitor
So for monitoring with Prometheus one could do something like
GET http://www.google.com" | vegeta attack -duration=5m | vegeta monitor -type='prometheus[8880,0.0.0.0]'
- "-type" may be omitted. defaults to 'prometheus'
- "8880" is the Prometheus exporter http port (may be omitted and defaults to 8880)
- "0.0.0.0" is the bind host for the listener (may be omitted and defaults to 0.0.0.0)
After launching the tests, just point Prometheus scraper to http://host:8880/metrics or do something like
curl http://host:8880/metrics
Prometheus Metrics
-
request_seconds_count{method, path, status} -
request_seconds_sum{method, path, status} -
request_bytes_sum{method, path, status} -
response_bytes_sum{method, path, status}
What do you think?
That’s a good idea.
Sent via Superhuman iOS ( https://sprh.mn/[email protected] )
On Tue, Jun 30 2020 at 1:46 AM, Flavio Stutz < [email protected] > wrote:
Proposal
What do you think about creating a new command category called "monitor" aside report, encode, dump?
Monitor
So for monitoring with Prometheus one could do something like
GET http://www.google.com" | vegeta attack -duration=5m | vegeta monitor -type='prometheus[8880,0.0.0.0]'
- "-type" may be omitted. defaults to 'prometheus'
- "8880" is the Prometheus exporter http port (may be omitted and defaults to 8880)
- "0.0.0.0" is the bind host for the listener (may be omitted and defaults to 0.0.0.0)
After launching the tests, just point Prometheus scraper to http://host:8880/metrics or do something like
curl http://host:8880/metrics
Prometheus Metrics
- request_seconds_count{method, path, status}
- request_seconds_sum{method, path, status}
What do you think?
— You are receiving this because you commented. Reply to this email directly, view it on GitHub ( https://github.com/tsenart/vegeta/issues/477#issuecomment-651426528 ) , or unsubscribe ( https://github.com/notifications/unsubscribe-auth/AAAQPDZZWGP2IH7P6UECLR3RZERU5ANCNFSM4KVEPIMA ).
Alternatively we’d instrument attack directly.
Sent via Superhuman iOS ( https://sprh.mn/[email protected] )
On Tue, Jun 30 2020 at 8:31 AM, < [email protected] > wrote:
That’s a good idea.
Sent via Superhuman iOS ( https://sprh.mn/[email protected] )
On Tue, Jun 30 2020 at 1:46 AM, Flavio Stutz < [email protected] > wrote:
Proposal
What do you think about creating a new command category called "monitor" aside report, encode, dump?
Monitor
So for monitoring with Prometheus one could do something like
GET http://www.google.com" | vegeta attack -duration=5m | vegeta monitor -type='prometheus[8880,0.0.0.0]'
- "-type" may be omitted. defaults to 'prometheus'
- "8880" is the Prometheus exporter http port (may be omitted and defaults to 8880)
- "0.0.0.0" is the bind host for the listener (may be omitted and defaults to 0.0.0.0)
After launching the tests, just point Prometheus scraper to http://host:8880/metrics or do something like
curl http://host:8880/metrics
Prometheus Metrics
- request_seconds_count{method, path, status}
- request_seconds_sum{method, path, status}
What do you think?
— You are receiving this because you commented. Reply to this email directly, view it on GitHub ( https://github.com/tsenart/vegeta/issues/477#issuecomment-651426528 ) , or unsubscribe ( https://github.com/notifications/unsubscribe-auth/AAAQPDZZWGP2IH7P6UECLR3RZERU5ANCNFSM4KVEPIMA ).
Much better!
Instrumenting “attack” is great because it avoids parsing stdin contents and performance will be better for large scale distributed deployments.
Sent from my iPhone
On 30 Jun 2020, at 03:32, Tomás Senart [email protected] wrote:
Alternatively we’d instrument attack directly.
Sent via Superhuman iOS ( https://sprh.mn/[email protected] )
On Tue, Jun 30 2020 at 8:31 AM, < [email protected] > wrote:
That’s a good idea.
Sent via Superhuman iOS ( https://sprh.mn/[email protected] )
On Tue, Jun 30 2020 at 1:46 AM, Flavio Stutz < [email protected] > wrote:
Proposal
What do you think about creating a new command category called "monitor" aside report, encode, dump?
Monitor
So for monitoring with Prometheus one could do something like
GET http://www.google.com" | vegeta attack -duration=5m | vegeta monitor -type='prometheus[8880,0.0.0.0]'
- "-type" may be omitted. defaults to 'prometheus'
- "8880" is the Prometheus exporter http port (may be omitted and defaults to 8880)
- "0.0.0.0" is the bind host for the listener (may be omitted and defaults to 0.0.0.0)
After launching the tests, just point Prometheus scraper to http://host:8880/metrics or do something like
curl http://host:8880/metrics
Prometheus Metrics
- request_seconds_count{method, path, status}
- request_seconds_sum{method, path, status}
What do you think?
— You are receiving this because you commented. Reply to this email directly, view it on GitHub ( https://github.com/tsenart/vegeta/issues/477#issuecomment-651426528 ) , or unsubscribe ( https://github.com/notifications/unsubscribe-auth/AAAQPDZZWGP2IH7P6UECLR3RZERU5ANCNFSM4KVEPIMA ).
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or unsubscribe.
Sure! :)
On Wed, Jul 1, 2020 at 6:12 AM Tomás Senart [email protected] wrote:
@flaviostutz https://github.com/flaviostutz: Contributions welcome! :)
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/tsenart/vegeta/issues/477#issuecomment-652297362, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB3N4XFQJ6DQGRNBYCMXLD3RZL4WXANCNFSM4KVEPIMA .
Is there any update on the status of this PR? This feature would be really useful, thanks!
Is there any update on the status of this PR? This is a very useful feature in our DevOps process.
Hi!!
Can you please add a command example on the docs? I've been trying to make it work but simply cant.
