Supabase launches OpenTelemetry traces and metrics in .NET SDK
Supabase introduced OpenTelemetry tracing and metrics support in its .NET SDK, covering Auth, Postgrest, Storage, and Functions clients. This enables integrated telemetry without adding OpenTelemetry dependencies or overhead when unused.
Before · Aug 15, 2026—
AfterThe .NET SDK emit traces and metrics through System.Diagnostics. Each client exposes its own ActivitySource and Meter, so you subscribe to them the same way you subscribe to any other .NET instrumentation.
Why it matters. Competitors should note Supabase's enhanced observability features in their SDK, improving developer experience and monitoring capabilities without extra dependencies.
Fly.io Introduces New Pricing Structure for Machines, Sprites, and Managed Postgres
Fly.io has replaced its 'Pay-as-You-Go Infrastructure' and 'Public Cloud Billing' with detailed, itemized pricing for Machines, Sprites, and Managed Postgres services. The new structure includes specific per-second and per-hour rates for compute, per-GB/month for storage, and tiered egress pricing, along with new plans for Managed Postgres clusters.
Before · Sep 1, 2026Plans get complicated, so we just charge based on usage. Pick and choose which pieces you need for your application; that's all you'll see on your invoice.
AfterHere's what it costs
Machines
Sprites
Managed Postgres
Why it matters. This is a significant shift from a usage-based, simplified billing model to a more granular, itemized pricing structure. Competitors should note the introduction of specific pricing tiers for different services, which could impact cost comparisons and competitive positioning.
Render updates pricing, storage, RAM, and connections
Render has adjusted its pricing structure, introducing new storage tiers and modifying RAM and connection limits. The company also clarified its policy on migration professional services.
Before · Sep 1, 2026Scale up to 40 GB of RAM, 1,000 connections, $32/month
AfterScale up to 50 GB of RAM, 3000 connections, $20/month, 256 MB storage
Why it matters. Competitors should note Render's updated pricing and increased limits, which could impact their competitive positioning. The clarification on migration services might also influence customer acquisition strategies.
Railway Launches New Sandbox & VM Offering with Detailed Pricing
Railway has introduced a new 'Sandboxes & VMs' offering, providing isolated Linux VMs with usage-based billing for CPU and memory. This new service includes various sandbox sizes, idle timeout options, and features like checkpoints and private networking.
Before · Sep 5, 2026Enterprise Support, Cloud Regions, Higher Resource Limits
Everything above, plus prioritized support with SLOs, a dedicated Slack Connect channel, cloud regions, and services with 48vCPU/48GB RAM per replica.
AfterSandboxes & VMs
Isolated Linux VMs for agents, builds, and untrusted code. Billed only for the CPU your sandbox actually uses and the memory it holds while running, never for the time it spends waiting.
Learn more
$ 0.00001929 per GB / sec
about $ 50 per GB-month, while running
$ 0.00001929 per vCPU / sec
about $ 50 per vCPU-month, active use only
for sandboxes
Sandboxes
Running sandboxes per environment
Default sandbox size
2 vCPU / 2 GB
4 vCPU / 4 GB
8 vCPU / 8 GB
Maximum sandbox size
2 vCPU / 2 GB
8 vCPU / 8 GB
32 vCPU / 32 GB
Idle timeout
Up to 5 minutes
Up to 2 hours, or never
Checkpoints, forks, and templates
Private networking and public domains
Enterprise Support, Higher Resource Limits
Everything above, plus prioritized support with SLOs, a dedicated Slack Connect channel, and services with 48vCPU/48GB RAM per replica.
How are sandboxes billed?
Sandboxes run on Railway's VM primitive and bill on VM rates: $50 per GB of memory per month and $50 per vCPU per month, metered per second, plus $0.05 per GB of egress. You pay for the CPU a sandbox actually uses and the memory it holds while running, not for wall-clock uptime, so a sandbox waiting on a model or a user costs almost nothing. Destroyed sandboxes cost nothing, and an idle timeout destroys forgotten ones automatically. Sandbox usage draws from the same included credit as the rest of your plan.
Why it matters. Competitors should care as Railway is expanding its platform capabilities with a new compute primitive, potentially attracting developers needing isolated environments for agents, builds, or untrusted code. The detailed, usage-based pricing model for VMs could also influence competitive pricing strategies.
Neon has increased the free storage per project from 0.5 GB to 1 GB and removed the 'Beta' label from Object Storage, Functions, and AI Gateway, indicating general availability for these services. A new 'Realtime (Beta)' service has been introduced, with its usage being free during the beta phase.
Before · Sep 5, 20260.5 GB of storage per project
After1 GB of storage per project
Why it matters. This change directly impacts the free tier offering, making it more generous for storage. The removal of beta labels suggests these services are now considered stable and ready for production use, while the introduction of Realtime (Beta) indicates a new product offering that could expand Neon's capabilities.
Supabase updates compute offerings and adds new features
Supabase has updated its compute offerings, changing from ARM CPU to vCPUs and introducing 'Shared compute' for some tiers. The company also added several new features and clarifications, including 'Technical Account Management', 'Supports AWS PrivateLink', 'Log Ingestion', 'Log Query', and 'OAuth 2.1 Server'.
Before · Aug 22, 20261 GB RAM / 2-core ARM CPU / Connections: Direct 60 , Pooler 200
Why it matters. A competitor should care about these changes as they indicate a shift in how Supabase packages and describes its compute resources, potentially impacting performance expectations and pricing comparisons. The addition of new features like AWS PrivateLink support and enhanced logging capabilities could also make Supabase more attractive to certain enterprise customers.
Netlify adds 'Concurrent agent runs' to pricing page
Netlify has introduced 'Concurrent agent runs' as a new metric on its pricing page. This metric defines how many agent runs a team can have active simultaneously across all projects.
Before · Sep 5, 2026—
AfterConcurrent agent runs
How many agent runs your team can have active at the same time, across all of your projects.
Why it matters. This change indicates a new dimension for how Netlify may be metering or limiting usage, potentially impacting CI/CD or build concurrency for teams. Competitors should monitor if this becomes a new lever for plan differentiation or a source of overage charges.
Vercel updates deployment limits and AI Gateway offerings
Vercel has updated its policy regarding monorepo deployments, where skipped deployments no longer count against rate limits. Additionally, new AI models have been made available on the AI Gateway, and WebSocket support is now generally available.
Before · Sep 4, 2026—
AfterSkipped monorepo deployments no longer count against limits
Why it matters. Competitors should note the change in deployment limit calculation, which could make Vercel more appealing for monorepo users. The expansion of AI Gateway models and GA of WebSocket support indicate continued investment in AI and real-time application capabilities.
Vercel introduces Flat Rate CDN and new 'Drives' and 'Daemons' services
Vercel has introduced a new 'Flat Rate CDN' pricing model, replacing usage-based billing for CDN requests and data transfer. Additionally, Vercel has launched new services called 'Vercel Drives' for storage and 'Vercel Daemons' for active CPU and memory, each with their own pricing and included limits.
Before · Sep 4, 2026Edge Requests 10M / month included then starting at $2 per 1M; 1TB / month included then starting at $0.15 per GB
AfterFlat Rate CDN with traffic spike protection; CDN Requests Included in Flat Rate CDN; Vercel Drives 15 GB (lifetime) included Starting at $0.05 per GB-month; Vercel Daemons
Why it matters. Competitors should care as Vercel is shifting its CDN pricing to a flat-rate model, which could appeal to customers with unpredictable traffic spikes and offer more predictable billing. The introduction of 'Drives' and 'Daemons' indicates an expansion of Vercel's platform capabilities into storage and background processing, potentially increasing its competitive scope.