Cloudflare Meerkat adopts QuePaxa after 2020 leader stall
First merged summary.
Cloudflare’s experimental Meerkat control-plane service uses leaderless QuePaxa after a 2020 leader-based stall.
Cloudflare says experimental Meerkat, which keeps an ordered log for an internal global control plane, uses QuePaxa instead of Raft, Multi-Paxos, or Viewstamped Replication. Those leader-based protocols can stall on wide-area links when the leader is slow or disconnected even if replicas have not crashed. Cloudflare cites a 2020 switch failure that left a consensus-backed system unable to take writes and about 80 times slower for six hours, without data loss. QuePaxa is meant to keep progressing without a designated leader. Meerkat’s log can back a key-value store; named comparable systems are Consul, etcd, and Apache ZooKeeper.
- Meerkat is an experimental internal global control-plane consensus service.
- QuePaxa replaces Raft, Multi-Paxos, and Viewstamped Replication.
- A 2020 switch failure blocked writes for six hours without data loss.
- The affected system was about 80 times slower during the incident.
- Meerkat replicas agree on an ordered log used for gets and puts.
- Comparable systems named are Consul, etcd, and Apache ZooKeeper.
Timeline
dated events in the reporting · oldest first- Nov 11, 2026Cloudflare explains Meerkat’s choice of QuePaxa · Cloudflare Blog
What each outlet adds
beyond the shared facts- Cloudflare Blog2020 six-hour write stall and leaderless QuePaxa rationale
- Vendors
- Cloudflare
- Organizations
- Cloudflare
Coverage
oldest first · one row per article · UTC- · 6h ago
Cloudflare BlogFirst report
Cloudflare chooses leaderless QuePaxa for its Meerkat control plane
A 2020 partial switch failure left a leader-based Cloudflare system unable to take writes and about 80 times slower for six hours.
- · 6h ago
Cloudflare picks leaderless QuePaxa consensus for Meerkat control plane
Meerkat needs progress across continents when latency spikes, which leader-based Raft and Paxos do not guarantee.