[gcp_authn] Support for ahead-of-time fetching of tokens
- Dominant language
- C++
- Stars
- 28.9k
- Forks
- 5.6k
- Avg merge
- 1d 20h
- Merged PRs (30d)
- 437
Description
*Title*: Have gcp_authn filter support ahead-of-time fetching of tokens
*Description*:
Right now, the filter supports fetching tokens in an "on-demand" fashion, with the value for audience
being the URL of the destination service. Thus, the first incoming request that will be routed to
the destination service will incur the extra cost of fetching the token from the GCE metadata server.
Similarly, the first incoming request to the destination service after token expiration will incur the extra cost
of fetching a new token.
In our use case, we have a set of services for which we need to fetch tokens, and we would like
to have these tokens fetched "ahead-of-time", so that requests routed to these services don't incur any
extra cost. Similarly, cached tokens that are about to expire would be replaced in the background.
Very curious to hear your thoughts about this use case @tyxia, and whether you think it would make sense
to have the filter support it :-)
Contributor guide
Assessment
This issue has not been assessed yet.