Pgn/stat metrics#101
Conversation
|
Warning You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again! |
| typedef struct tcp_rtt { | ||
| u8 flags; // Must be first, we use it to tell what kind of event we have on the ring buffer | ||
| u8 _pad[3]; | ||
| u32 srtt; |
There was a problem hiding this comment.
can we change the name of this var to also contain the measure type used? eg. srtt_us or srtt_ms
There was a problem hiding this comment.
right! I think srtt_ms is good, I changed also the go struct field to SrttMs even if at the end is exported as seconds.. maybe we can decide the best option
Ok I think it's better to use srtt_us for the entire path and then at the end export it in second.
|
|
||
| func log() *slog.Logger { return slog.With("component", "k8s.MetadataDecorator") } | ||
|
|
||
| func MetadataDecoratorProvider( |
There was a problem hiding this comment.
I see now why this code is here, I think se should define the types first, and refactor all these enrichers to be generic, maybe in a different PR
There was a problem hiding this comment.
I was about to write this in the previous comment.. it seems too much to modify it in this PR but yes, this and other decorators could be common!
| // TODO pinoOgni: right now it is a copy and paste of more or less netolly config | ||
| type StatsConfig struct { | ||
| // AgentIP allows overriding the reported Agent IP address on each stat. | ||
| AgentIP string `yaml:"agent_ip" env:"OTEL_EBPF_STATS_AGENT_IP" validate:"omitempty,ip" jsonschema:"type=string,format=ip"` |
There was a problem hiding this comment.
should this be refactored to inherit from some struct in common with netolly?
There was a problem hiding this comment.
I was thinking something similar but in that case we will have issues with the env variables
|
|
||
| GeoIP stats.GeoIP `yaml:"geo_ip"` | ||
|
|
||
| // TODO pinoOgni add more granular configuration |
There was a problem hiding this comment.
I would like to add a more granular configuration, not now but in future PRs so the user can select which metric calculate. Right now the idea is to "calculate everything and then the user can filter using attributes or the collector"
WDYT?
Update: I removed the comment, maybe I can do it in another PR
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
exported with prom and otel Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
add iters in the appnetworktracer Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Pin listening_ports map. Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
probes to calculate metric if the hook point is equal Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
as stats metrics Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
- Define statsolly - Define statsolly pipeline - Add sample tcp rtt ebpf probe - Update configuration to use statsolly Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
…olly Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
…ostInfo Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
35f3892 to
bb8d207
Compare
Signed-off-by: Giuseppe Ognibene <giuseppe.ognibene@coralogix.com>
96cd0bc to
009314b
Compare
|
I think we should split the PR in multiple smaller ones, something like:
I think it would be a bit too much to review all three in a single PR upstream |
Uh oh!
There was an error while loading. Please reload this page.