Hey everyone,
Before today nectarengine relied on a single, hard-coded URL for its API calls, or one you provided. If that one node was slow, offline, or just having a bad day, everything would time out and crash.
That's a terrible design, I needed to fix it properly.
NectarEngine 0.1.4: Now with Less Failure
I'm happy to announce the release of NectarEngine 0.1.4, which is a major refactor of how the library handles node connections.
Instead of a single RPC instance, I've built a new _RPCPool that is much more resilient. Now, when you initialize the Api, it creates a pool of candidate nodes. If a call fails, it automatically retries a couple of times and then rotates to the next node in the list. This should make the bot, and any app using the library, far more reliable.
Smarter Node Discovery
I didn't just want to replace one hard-coded list with another. The library is now much smarter about finding good nodes.
- It automatically pulls the latest, best-ranked nodes from the PeakD Beacon (
https://beacon.peakd.com/api/he/nodes). - It also uses the
@flowerenginemetadata as a fallback. - If both of those fail, it falls back to the old hard-coded default.
This new logic also applies to Hive Engine history nodes. The library now pulls a list of history nodes from the beacon, and it will retry up to ten times on a failed query before giving up.
Housekeeping
While I was in there, I did a lot of spring cleaning. I stripped out a ton of old, unused __future__ import statements and other legacy code. I also expanded the test suite to cover all the new pool rotation and retry logic, ensuring the local src/ package is properly imported during testing.
The new version is all built, tested, and published to PyPI.
Hopefully, this means you'll be seeing a lot less of that type of error.
As always,
Michael Garcia a.k.a. TheCrazyGM
Nice update.
Love to see Beacon integrated into more and more project 😍
Thank you for updating this!
!PAKX
!PIZZA
!PIMP
View or trade
PAKXtokens.Use !PAKX command if you hold enough balance to call for a @pakx vote on worthy posts! More details available on PAKX Blog.
$PIZZA slices delivered:
@ecoinstant(1/20) tipped @thecrazygm
Come get MOONed!
Excellent, my friend! This will certainly help make infrastructure that depends on functional nodes a good bit more resilient, and with all the intense attacks on Hive and Hive-Engine nodes, this is essential now I'd say! 😁🙏💚✨🤙
Nice update, keep forgetting to test it. Been using my own library for a while now because I wanted smooth fail over retry logic, plus everything else was abandoned.
Excellent project.