Degraded networking to AWS IPs
This incident has been resolved, connectivity from AWS to all of our hosts is back to normal.
The latest update from Fly.io.
State
Resolved
Started
6 d ago
2026-10-26 14:34 UTC
Lasted
6 h 6 min
until 2026-10-26 20:40 UTC
Worst impact
Notice
Timeline
7 updatesInvestigating
41 min 11%
Identified
5 h 25 min 89%
Updates
Times in UTCAs posted by Fly.io
This incident has been resolved, connectivity from AWS to all of our hosts is back to normal.
Connectivity from AWS to Fly.io on impacted hosts is continuing to recover. Approximately 80% of the hosts that were seeing connectivity issues are fully recovered at this time.
AWS has implemented changes to unblock connections from their network from reaching our machines. We are seeing connectivity starting to resume on a host-by-host basis as these changes take effect.
We are continuing to work with our upstream networking partners and AWS on remediation. Impacted customers are recommended to allocate app-scoped static egress IPs using **fly ips allocate-egress.** Customers seeing issues are encouraged to also contact AWS support or their AWS hosted vendors support with details of the connection errors and their fly machine IPs.
We are continuing to see connectivity issues from AWS IPs to a subset of our hosts. From our investigation these connections appear to be being blocked from the AWS side. We are continuing to work with our upstream networking partners and AWS on remediation. In the meantime impacted customers are recommended to allocate app-scoped static egress IPs using **fly ips allocate-egress.**
Our initial investigations indicate connection errors are occuring the return path from AWS IPs. We are escalating the issue with AWS and continuing to investigate. Customers seeing issues are encouraged to also contact AWS support or their AWS hosted vendors support with details of the connection errors and their fly machine IPs.
We are investigating reports of elevated networking issues between our network and AWS IPs. We are working to determine the source of the issue.
Fly.io in the last 90 days
Full historyIncidents in 90 days
19
Days with an incident
15
Median duration
1 h 38 min
This incident
3.7× the median
Other Fly.io incidents
All incidentsFly.io·Slow image pulls delaying Machine starts in IAD
Image pull times in IAD have returned to normal.
Fly.io·Elevated issues with the Machines API
The incident has been resolved.
Fly.io·Creating WireGuard peers in ARN is currently broken
The ARN WireGuard gateways are up
Fly.io·Packet loss in IAD
The congestion affecting our upstream provider has been resolved. We are observing a return to normal levels of networking performance in IAD region.
Fly.io·MPGv2 IAD control plane is read-only
Control plane operations in IAD, including resizing and changing plans, are working normally again.
Follow Fly.io
Your coding agent knows when Fly.io is down, and tells you when a failure is theirs, not yours. You can also get Fly.io incidents in a feed reader or in Slack.
Install the outages.live MCP server, so that when something breaks you can tell if the cause is a provider we depend on and not my code. It reads the official status pages of cloud, AI, payment and developer providers and gives you their incidents as MCP tools.
Add it as a remote MCP server named "outages" at https://outages.cc/mcp (Streamable HTTP, no authentication), with your CLI command or your MCP config file.
Then:
1. Make sure it connects and shows the tools get_provider, list_feed and send_feedback.
2. Call get_provider with "fly" and tell me if Fly.io has a problem right now.
3. From now on, when a call to a third-party service fails or becomes slow, call get_provider for that service before you change my code.Paste this into any coding agent. It adds the server with its own MCP setup.
https://outages.cc/feed.xml?providers=fly/feed subscribe https://outages.cc/feed.xml?providers=flyFAQ
Go from tracking outages to fixing production issues with Polylane.
Try Polylane