What is the maximum number of rpc requests multichain can handle per second?

+1 vote

After facing issues with slow rpc responses from the multichain daemon (previously asked https://www.multichain.com/qa/5558/why-are-the-rpc-calls-getting-slow), i changed the machine to a new 2gb RAM Virtual machine with the previous data. I faced the same issues here as well. It seemed even slower if the explorer window was running in the browser (we integrated a previous bitcoin explorer by changing the rpc calls). The page has a setTimeout command hitting an end point on the back end which makes 10 rpc calls every 5 second(120 per minute). There are two scripts populating the explorer db. One hits 6 rpcs per minute and the other one hits 1 rpcs every 5 seconds(12 rpcs/min). The transaction generating script has a hit rate of 6/minute. All in all 144rpcs/minute. 

The new setup is such that the multichain server is running on an isolated server and all explorer based programs (server engine, database, etc) on a separate server only communicating with the multichain server over rpcs.

The same setup for a bitcoin explorer works pretty smoothly, even when all the applications are running , it instantly returns a value(the rpc).
I am confused as to where the problem lies in this case. Please help me resolve this issue.

asked Jul 10, 2017 by sma1692

1 Answer

0 votes

As we mentioned on the previous post, we couldn't see a problem with API throughput when we brought your blockchain up on a regular low-end server. So the most likely explanation is that something is hitting the API hard with requests that take a long time to respond to.

To help identify if this is the case, please stop MultiChain then run it as usual with the parameter -debug=mcapi added to the command line. That will add mcapi lines to debug.log to show the API requests received and when they are completed. If you post all the matching lines from a one minute period here, we can take a look.

answered Jul 10, 2017 by MultiChain