Fix namespace valid chars matching regex #6352
Conversation
| @@ -274,10 +274,16 @@ def collect_latency(self, namespaces): | |||
| continue | |||
|
|
|||
| # match only works at the beginning | |||
| ns_metric_name_match = re.match(r'{(\w+)}-(\w+):', line) | |||
| # ':' or ';' are not allowed in namespace-name: https://www.aerospike.com/docs/guide/limitations.html | |||
| ns_metric_name_match = re.match(r'{([^\}:;]+)}-(\w+):', line) | |||
There was a problem hiding this comment.
Awesome. Could you take the opportunity to compile the regex in init?
There was a problem hiding this comment.
@therve Regexes are cached by python, not sure it worth moving them away from where they are used.
There was a problem hiding this comment.
Not really in a reliable fashion though. If you have another integration in the same agent uses lots of regexp, then your cache is gone.
I think moving them a bit but giving them a readable name makes sense.
There was a problem hiding this comment.
§ python2 -c 'import re; print(re._MAXCACHE)'
100
§ python3 -c 'import re; print(re._MAXCACHE)'
512
The regex cache is quite small. Yeah, probably better to not rely on the cache.
If you don't mind, I will do a refactoring to pre compile all the regexes in a separate PR. We might want to cherry pick this PR (probably not, just in case).
cc @ian28223 @therve
What does this PR do?
Fix namespace valid chars matching regex
Namespace might contains chars like
-: https://www.aerospike.com/docs/guide/limitations.htmlReview checklist (to be filled by reviewers)
changelog/andintegration/labels attached