Skip to content

Potential request spamming after initial cooldown surpassed #137

Closed
@earloc

Description

@earloc

Hey Folks.

We recently had a situation in one of our environments, where usage of the refresh-api somehow overwhelmed our azconfig instances, which then gave HTTP 429 Too-many-requests, occasionally - especially during peak-load times of our app.

We used 1.0.0-preview4 at this time and thx to @lorenzadomat found out, that the refresh-interval might be ignored, somehow.

I then took the opportunity to narrow this even further down. Seems like starting with 1.1.0 / b1cf8099fbf8737de3d38299be2871d4547c7cac , this behavior dissappeared.

Looks like prior 1.1.0, sentinel keys got queried every time, after the initial refreshIntervalInMs has been surpassed - before this has happened, everything works as expected - that's why this was not that easy to reproduce ;)

Just wanted to let you know about this, especially as this might have been fixed "by accident" and not on purpose.

I have a simple repro over at my repro-repo, which hopefully showcases the problem clear enough.

Let me know if this somehow might help in preventing some future regressions. We will probably notice, at least.

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions