25 days ago
We're seeing latency spikes and intermittent dropouts since around the 6th or 7th of June across all deployments. We track it here (https://status.balladic.com/) if you're interested.
This seems to line up with the recent Hikari rollout. Is there something we need to handle on our end, or is this purely a Railway-side issue out of our hands?
Thanks in advance,
Adrian on behalf of Balladic
3 Replies
25 days ago
All seven services in your Balladic Core project show healthy deployments, and we don't have any documented platform-wide latency issues or incidents aligned with the June 6-7 timeframe. We also don't have internal context on a "Hikari rollout" to confirm or deny a correlation. To help us investigate further, could you share the X-Railway-Request-Id header from some of the slow or dropped requests? That lets us trace specific requests through our networking layer and determine whether the latency is on our side.
Status changed to Awaiting User Response Railway • 25 days ago
25 days ago
The Hikari rollout is written up here: https://blog.railway.com/p/railway-cdn. I'm speculating that this is related, it might not be at all. Deployment health is not in question, this is unrelated.
Here's some affected requests (seconds refering to TTFB):
zXhW10brQKS7nB8d63j4Ww = 4.69s
hAM2wOZrRgWbwyl_63j4Ww = 2.03s
yxXbH7vySwers9c163j4Ww = 2.01s
0CYf2_keQQWR0HoSEuroCg = 1.62s
yYwnUvSBQn2TeG3IEuroCg = 1.01s
Status changed to Awaiting Railway Response Railway • 25 days ago
22 days ago
We've been rolling out some changes that should be fixing this. Just want to check are you still seeing these issues? Any recent ids to trace with?
Status changed to Awaiting User Response Railway • 22 days ago