RDPForge network status — all systems operational

1. DegradedMay 4, 2026· 38 minutes

   Elevated packet loss — Warsaw

   A transit provider's border router in Warsaw began dropping ~4% of packets on one of three uplinks after a line-card fault. We depreferred the path via BGP local-pref within 6 minutes of detection; the provider replaced the card 32 minutes later. Customer impact was intermittent latency, not loss of connectivity. No SLA credits were triggered.

2. OutageNovember 19, 2025· 22 minutes

   Client area unavailable

   A database migration during a routine deploy took an exclusive lock on the invoices table, blocking every client-area request. Servers were entirely unaffected — only billing, ticketing and the console were down. Migrations are now gated behind a lock-duration estimate that aborts anything projected over two seconds.

3. DegradedJuly 2, 2025· 3 h 12 m

   Elevated write latency — Singapore

   A storage controller firmware revision introduced a write-amplification regression on four nodes, pushing p99 write latency from 0.7 ms to 14 ms. Not an outage, but customers noticed and were right to. We rolled the firmware back and now stage all storage firmware for 30 days on internal nodes before production rollout. SLA credits were applied automatically to the 214 affected instances.

4. OutageMarch 14, 2025· 47 minutes

   Partial reachability loss — Frankfurt

   During planned maintenance our upstream failed to withdraw a route, blackholing roughly 30% of Frankfurt-bound traffic. Our monitoring detected it in 90 seconds; the provider took 44 minutes to roll back. We have since added a third transit provider in Frankfurt and reduced BFD timers to 300 ms, so an equivalent event now fails away in under a second. Full credits applied.

5. DegradedDecember 8, 2024· 1 h 05 m

   Delayed provisioning queue

   A Black Friday order backlog exceeded provisioner concurrency and average delivery rose from 4 minutes to about 26. No orders failed; all completed. The provisioner was rewritten in Q1 2025 to parallelise image write, IP assignment and DNS registration.

6. DegradedJune 21, 2024· 11 minutes

   Volumetric attack absorbed — Amsterdam

   A 1.14 Tbps UDP reflection attack targeting a single customer IP. Edge scrubbing engaged in 9 seconds; the target and neighbouring instances stayed reachable throughout, with brief latency elevation on the Amsterdam edge. No null-route was required. Published here because absorbing an attack quietly still counts as an event.

7. OutageNovember 2, 2023· 26 minutes

   Hypervisor kernel panic — Dallas node 41

   A kernel bug in the 5.15 LTS branch panicked one hypervisor, taking 38 instances offline. All were restored from the same node after reboot with no data loss. We moved the fleet to a patched kernel over the following week. This is the first incident we published in full; we have published every one since.

SLA credits are applied automatically when monthly uptime for an instance falls below 99.9% — you never have to ask. See the service level agreement.