GetBlock RPC competitor monitor — 2026-11-04
Criterion: RPC performance, latency, reliability, SLA, uptime, regional routing and failover claims (rpc-performance-reliability)
Decision summary
- Confirmed decision-relevant market changes: 0. Several monitored hashes changed, but the captured differences were dominated by navigation, campaign banners, chain-list updates, or blog churn. These are not treated as product changes.
- The cumulative topic brief was refreshed to 2026-11-04 using direct vendor evidence while preserving previously verified evidence where today's capture did not supersede it.
- Strongest current public specificity in this pass: QuickNode publishes a 99.99% enterprise uptime commitment and service-credit wording; Ankr publishes 56 ms average response time, 99.99% uptime and geo-distribution across 24 metros; Lava publishes 99.99% uptime plus real-time routing and automatic provider failover.
- Volatile performance, uptime and SLA statements must be verified in the buyer's service agreement and workload context before sales use.
Evidence assessment
Stable direct evidence refreshed
- QuickNode: home states 99.99% uptime, low-latency RPC, dedicated clusters and up to 50,000 RPS burst; enterprise states a contractual 99.99% uptime commitment, service credits, measurement across production endpoints, and a 2.5× benchmark claim. The pricing page confirms contractual uptime SLAs for Enterprise. These claims were stable versus the prior snapshot.
- Infura: home continues to claim high availability, reliability, speed and throughput; pricing lists an Enterprise enhanced-support SLA and auto-scaling, but the public capture does not state an uptime percentage or service-credit formula.
- Chainstack: home continues to describe Global Nodes as geo-balanced and ultra-reliable and announces GEN 2.0 lower latency. The changed blog hash did not establish a distinct product or contractual change.
- Ankr: home states 56 ms average response time and 99.99% uptime. Enterprise states geo-distribution across 24 metros and customizable throughput above 25,000 RPS. Changes on RPC/docs pages did not provide enough evidence to classify a new reliability capability.
- NOWNodes: home states greater than 99.95% API uptime and approximately 0.2-second response time. The home-page hash changed, but the capture does not prove a new performance or reliability offer.
- dRPC: home describes multichain endpoints with AI load balancing, eight geo-distributed clusters and low latency. The changed chain-list capture reflects coverage/list content, not a verified reliability change.
- Lava Network: home states 99.99% uptime, live provider scoring/routing and automatic failover across independent providers. No page-level change was detected.
Changed captures classified as noise or insufficient evidence
- Alchemy: home, pricing, enterprise and blog hashes changed. Captures prominently repeat campaign banners and shared navigation. The pages still market low-latency RPC/gRPC, dedicated clusters and “world-class reliability and performance,” but the supplied extraction does not expose a quantitative uptime/SLA, regional-routing or failover commitment. No market change confirmed.
- Chainstack: blog changed; the sample is dominated by repeated site navigation and GEN 2.0 promotion already visible elsewhere. No distinct market change confirmed.
- Ankr: RPC and docs changed, but the extracted text shows plan/docs navigation and chain/tooling descriptions rather than a new measurable reliability commitment. No market change confirmed.
- NOWNodes: home changed while continuing previously observed uptime/response-time marketing. No new claim confirmed.
- dRPC: chain list changed, consistent with network-list churn. No reliability change confirmed.
- Blockdaemon: blog changed, but the capture is mostly global navigation. The home page confirms Native RPC and dedicated nodes, not a public latency, uptime, regional-routing or failover metric in this pass. No reliability change confirmed.
Provider gaps and service-state notes
- Moralis: current captures use general “low latency and reliability” positioning, without a public uptime percentage, routing topology or failover mechanism in the supplied evidence.
- Tatum: Intelligent RPC Routing is publicly named on home, but the supplied capture does not substantiate routing regions, failover behavior, latency or SLA terms.
- Blast API: official notice says Blast products are deprecated and directs users to Alchemy. Active Blast RPC performance comparison is therefore not applicable.
- Pocket Network / Grove: Grove says Grove is winding down and that Pocket Network continues under the Pocket Network Foundation. The monitored Pocket Network capture is now focused on agent data services and does not verify current managed RPC SLA/performance terms.
- LlamaRPC: home and chain pages returned HTTP 526; docs and status hostnames did not resolve. Claims remain Not verified in monitored public evidence; technical fetch failure is not interpreted as product absence.
- Allnodes: all five monitored URLs returned HTTP 403. Previously confirmed facts remain historical evidence, but no current 2026-11-04 performance/SLA claim could be refreshed.
Retrieval and extraction limitations
- QuickNode changelog: HTTP 404.
- Alchemy legacy docs-chain and customer-story paths: HTTP 404; valid current pages remained reachable.
- NOWNodes Postman documentation path: HTTP 404.
- Blockdaemon legacy RPC and pricing paths: HTTP 404.
- Tatum legacy supported-blockchains path: HTTP 404.
- Lava legacy developers and network paths: HTTP 404.
- LlamaRPC: HTTP 526 and DNS failures as described above.
- Allnodes: HTTP 403 across monitored surfaces.
- Generic marketing phrases such as “reliable,” “fast,” or “enterprise-grade” were not converted into quantitative or contractual claims.
GetBlock implications
- Publish workload-specific latency percentiles and methodology rather than an unqualified “fast” claim.
- Make contractual uptime, measurement scope, exclusions and service-credit schedule easy to find; QuickNode is unusually explicit publicly.
- Explain regional routing and failover separately. Geo-distribution alone does not prove automatic failover, and provider diversity is a distinct reliability control.
- Differentiate dedicated/single-tenant capacity from shared endpoint reliability and document burst behavior.
- Keep a public status/history surface and expose region- or chain-level health where possible.
Agentic routing paper
agentic_routing_paper_update.ok was true. The updater found 57 evidence items, updated and republished the paper, and reported paper_updated: true.
- Public paper: Agentic routing RPC
- Review areas recorded by the updater: cost/SLA constraints, enterprise-SLA metrics, method-specific routing, metrics and incident simulation.
Publication status
The pre-run encyclopedia publication reported success (published_urls: 149, latest daily date 2026-11-04). This report and the cumulative topic brief were subsequently consolidated and require the post-write publisher/HTTPS checks recorded by the job completion result.