Changing our RPC system in staging


#1

Boulder, the server software that powers Let’s Encrypt, has recently added support for a new internal RPC system which should simplify large parts of our codebase (and hopefully provide a small performance boost).

We are currently testing this new system in staging. Since it completely replaces the communication mechanism between all of our internal services we are testing this change more thoroughly than usual to try to root out any bugs which may have crept in. That said there is only so much we can look out for so we are asking that anyone running their own tests against staging keep their eyes out for any strange things they may come across, such as internal server (500) errors or responses missing expected fields etc and report them in this thread.

If you’d like to report something you should provide as much context as possible such as:

  • Time the behavior occured
  • Actions that led up to the behavior
  • Client software used (including version/revision)

Thanks for helping make Let’s Encrypt get better!


500s from staging server, now acme:error:rateLimited
#2

For the curious Boulder is moving from using RabbitMQ for RPC to using gRPC.


#3

We’ve now fully switched over our production systems to gRPC. Please let us know if you experience any problems.


#4

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.