On this page

Resource Limits

Every Guara Cloud plan comes with its own set of resource limits. These limits define how much compute, storage, and other resources you can allocate across your projects and services.

When you reach a plan’s limit, your existing services continue running normally — the platform simply prevents new allocations that would exceed your quota.

Limits by plan

This table is generated from the same plan catalog that powers the pricing page. Prices for each plan are in Plans & Pricing.

ResourceStarterProMaxBusinessEnterprise
Projects per account 99 99 99 99 Unlimited
Total vCPU (account pool) 1 vCPU 2 vCPU 4 vCPU 6 vCPU Custom
Total RAM (account pool) 2 GB 5 GB 12 GB 25 GB Custom
Total storage per project 20 GiB 30 GiB 50 GiB 100 GiB 500 GiB
Build minutes/month 500 1,000 2,000 4,000 20,000
Largest pod (chosen size) 1 vCPU / 2 GB 2 vCPU / 5 GB 4 vCPU / 12 GB 6 vCPU / 25 GB 4 vCPU / 8 GB
Automatic pod burst 0.5 vCPU / 256 MB 0.5 vCPU / 512 MB 1 vCPU / 512 MB 1 vCPU / 1 GB 4 vCPU / 4 GB
Replicas per service 10 10 10 10 10
Persistent volumes per project 5 10 15 20 100
Storage per volume 10 GiB 15 GiB 25 GiB 50 GiB 50 GiB
PostgreSQL recovery window 3 days 3 days 7 days 7 days 14 days
Log retention 14 days 14 days 14 days 14 days 14 days
Continuous profiling retention 3 days 7 days 7 days 7 days 7 days
Cron workers per project 3 10 10 10 Unlimited
Custom domains per service 10 10 10 10 10
Bandwidth Unlimited (fair use of 10,000 GB/mo) Unlimited (fair use of 10,000 GB/mo) Unlimited (fair use of 10,000 GB/mo) Unlimited (fair use of 10,000 GB/mo) Unlimited (fair use of 10,000 GB/mo)
Overage Optional Optional Optional Optional Included
Support Community Priority (12h) Priority (12h) Priority (8h) Dedicated (4h)

During the free trial, your account uses the Starter limits, except that overage cannot be enabled until you pick a plan.

How resources are distributed

Each limit applies at a different scope:

ResourceScope
vCPU and RAMAccount total, shared by every service across all projects
Total storagePer project, adding up every volume in the project
Storage per volumePer persistent volume
Build minutesAccount total, per month
ReplicasPer service

Your plan’s vCPU and RAM are not a fixed amount per service: you can split the total freely between your services. What each service takes from the pool depends on its pod size:

  • Automatic (default): the service reserves only the platform minimum and can burst up to the automatic pod burst in the table above. It’s the simplest way to run several small services.
  • Chosen size: the service reserves exactly the CPU and memory you pick for each replica, up to the plan’s largest pod. Use it for services that need guaranteed resources, such as a busy database.

Databases and other catalog services also draw from the account pool, like any other service.

Services running continuously

Every service runs continuously, 24 hours a day, without sleeping when idle. As long as what your services reserve fits in your plan’s pool, everything is included in the monthly price: there is no per-hour, per-service, or per-database charge. Extra charges only happen if you enable overage.

For example, on the Pro plan a Django app with four services — web, PostgreSQL, a scheduler, and Mailpit — shares 2 vCPU and 5 GB of RAM in total. With the services on Automatic, each one can burst up to 0.5 vCPU / 512 MB. If you prefer guaranteed resources, pick a size for each service (for example, more memory for web and PostgreSQL and the minimum for the scheduler and Mailpit), as long as the total fits in the pool. PostgreSQL and the project’s other volumes share 30 GiB of storage.

Choosing pod size and replicas

The CPU and memory in your plan form an account pool shared by every service and replica across your projects. You decide how to spend it in each service’s Resources page: a single large service, or many small ones.

  • Pod size: keep Automatic (the platform reserves a small minimum and lets the pod burst up to the automatic limit above), or choose a CPU and memory size. A chosen size is reserved for each replica: memory is guaranteed and CPU can still burst up to the plan’s limit.
  • Replicas: run a fixed number of replicas, or turn on autoscaling with a minimum, a maximum, and CPU and memory targets. Autoscaling reserves its maximum in the pool, so the platform can always add the replicas you allowed.
  • Pool: everything a service reserves is its pod size × its replicas (or autoscaling maximum). Sizes that wouldn’t fit what’s left of your pool can’t be chosen.
  • Services with a persistent volume run a single replica.

You can also change sizing from the CLI with guara scale, for example guara scale --cpu 500m --memory 1Gi --replicas 2.

How limits are enforced

When you try to create a resource that would exceed your plan’s limits, the platform rejects the request with a clear error message. The error shows your current usage and the maximum allowed by your plan, so you always know exactly where you stand.

Limits only apply to new allocations. If you’re already using all the resources your plan allows, you can still manage and operate your existing services — you just won’t be able to create additional ones until you free up capacity or upgrade.

If you need more resources, you have two options:

  • Upgrade your plan to get higher limits across the board.
  • Enable overage (available on Starter, Pro, Max, and Business plans) to pay for usage beyond your plan’s included amounts.

Compute overage

Overage charges for usage beyond the amounts included in your plan. It’s available on the Starter, Pro, Max, and Business plans.

When overage is enabled, CPU, memory, storage, and build minutes used beyond what your plan includes are charged by measured usage at the end of the billing cycle — you only pay for what you use. Overage doesn’t grow your account pool: the pod sizes and replicas you reserve always stay within it.

ResourcePrice
vCPU beyond the pool R$ 4/vCPU-hour
RAM beyond the pool R$ 2/GiB-hour
Build minutes beyond the quota R$ 0.03/minute
Storage beyond the project total R$ 2.50/GiB-month

Database backups

Automatic backups are included on every plan, including the free trial, at no extra cost. They’re turned on automatically when you deploy a catalog service that supports backups — there’s nothing to configure.

PostgreSQL

Catalog PostgreSQL is a managed database with continuous backup: every change is copied off the cluster within minutes, and a full copy is taken every day. That lets you restore the database to any point in time within your plan’s recovery window:

ResourceStarterProMaxBusinessEnterprise
PostgreSQL recovery window 3 days 3 days 7 days 7 days 14 days

A restore creates a new database and never overwrites the original. See Managed PostgreSQL.

The PostGIS and TimescaleDB variants use the snapshot backups described below.

Other catalog services

Catalog serviceFrequencyRetention
MySQL, MongoDB, RabbitMQ, Meilisearch, Qdrant, PostgreSQL (PostGIS and TimescaleDB)Daily3 most recent snapshots
MinIO, n8n, WebhookXDaily7 most recent snapshots
  • How it works: once a day, at a fixed time assigned to each service, the platform takes a snapshot of the data volume. When a new snapshot is created, the oldest one beyond the retention is discarded.
  • Where they live: snapshots are stored alongside the volume on the platform itself. To keep copies longer or outside Guara Cloud, also run periodic database exports, for example with mysqldump.
  • Services without automatic backups: Redis, Valkey, NATS, and other cache and messaging services keep persistent storage but have no automatic snapshots. For your apps’ persistent volumes, run guara volumes backups <volume> to check whether snapshots are available.

Restoring a snapshot

  1. Open the service page and go to the Backups tab
  2. Pick a snapshot with status Ready
  3. Click Restore and confirm

From the CLI, list snapshots with guara services backups and restore with guara services restore --snapshot <name>.

During the restore, the service is stopped, the volume is reverted to the chosen snapshot, and the service starts again. The database is unavailable until the restore finishes.

Upgrading your plan

When you upgrade your plan, the new limits take effect as soon as the upgrade is paid. You can start creating additional resources right away.

If you downgrade, the change takes effect at the end of the current billing cycle. Your existing resources stay available until then, even if they exceed the lower plan’s limits.

When the downgrade takes effect, everything above the new plan’s limits is adjusted: services drop to a single replica and chosen sizes go back to Automatic, starting with the largest; excess projects and services are deleted, excess volumes are disabled, and excess cron workers are paused. The dashboard shows these changes before you confirm. See Manage subscription.

For more on plans, invoices, and payments, see the billing documentation.