Unstable Hermes API

Hi,

I’ve been calling the Pyth public Hermes API (https://hermes.pyth.network/v2/updates/price/latest) from GCP in Singapore.

It used to be stable, but lately response times are highly unstable, sometimes over 1 minute, often exceeding 10 seconds.

The frequency of receiving data from websocket is also low. Is the server still healthy/stable? Do you have any suggestions on better regions or alternate endpoints to reduce latency?

Errors continuously happened, not around specific time. I set a 10 sec hard limit timeout and found reaching this limit once / 5-minutes.

Sometimes, the Websocket also stops sending the latest price data. Public subscription too

The first recommendation is to use a third-party node provider as well. You can find them here: https://docs.pyth.network/price-feeds/api-instances-and-providers/hermes#node-providers

The second recommendation would be using these nodes in other clusters.

These are all the three:

  1. https://hermes-stable-cyan.dourolabs.app/
  2. https://hermes-stable-green.dourolabs.app/
  3. https://hermes-stable-yellow.dourolabs.app/