Working with Google Cloud?

Google Cloud giving you grief? We sort it out.

A Google Cloud account usually arrives in one of two states: handed over by someone who built it in a hurry, or quietly running up a bill nobody can fully explain. We trace what each service is doing, lock the permissions down, and leave you with something a colleague can pick up on Monday.

  • Since 2015Fixing technology
  • FreeEstimates
  • US teamReal engineers

gcp.log

still failing

What it is

Google Cloud at a glance

Google Cloud Platform is Google's public cloud, running the same underlying infrastructure that powers Search, YouTube and Gmail. It began in 2008 with App Engine and grew outward from there, adding BigQuery for analytics, Compute Engine for virtual machines, GKE for Kubernetes and Cloud Run for serverless containers.

Its product range covers compute through GCE, GKE, Cloud Run and Cloud Functions; storage through Cloud Storage and Persistent Disk; databases through Cloud SQL, Spanner and Firestore; data through BigQuery, Dataflow, Dataproc and Pub/Sub; identity through Cloud IAM; and AI through Vertex AI and Gemini. Teams pick it most often when their centre of gravity is data analytics or machine learning.

Type
Public cloud platform
First released
2008 with App Engine
License
Proprietary pay-as-you-go
Maintained by
Google LLC
Typically used for
Data platforms, Kubernetes workloads, Cloud Run microservices, BigQuery analytics, Firebase mobile back ends, ML pipelines

The honest version

Google Cloud pros and cons

No technology is all upside. This is where Google Cloud earns its place and where it tends to cause trouble.

Where Google Cloud is strong

  • BigQuery and Vertex AIThe data and ML story is genuinely strong here. BigQuery handles very large analytical queries without a cluster to babysit, and Vertex AI gives model training and serving under one roof.
  • Cloud Run is lovelyPush a container, get an HTTPS URL that scales to zero and bills by the second. For most web back ends it removes a whole category of plumbing.
  • Generous free tierSeveral services keep working free for small loads, which makes side projects and proofs of concept cheap to leave running.
  • Firebase slotted inFirebase Auth, Firestore and Hosting share the same identity and billing, so mobile and web front ends talk to GCP without a translation layer.
  • Good IAM primitivesRoles, conditions and service accounts give you fine-grained control once you learn the model. Workload Identity ties Kubernetes pods to cloud permissions properly.
  • Serious KubernetesGKE is run by the people who invented Kubernetes, and Autopilot clusters hide most of the node management. Anthos extends the same control plane across other clouds and on-prem.

Where Google Cloud bites

  • Documentation is unevenSome services have thorough guides; others leave you reading release notes and GitHub issues. Which experience you get depends on the product.
  • IAM is rich but confusingBasic roles, predefined roles, custom roles, service accounts and Workload Identity all overlap. Teams routinely grant Owner just to stop fighting it.
  • Billing is easy to orphanA project without a linked billing account quietly stops working. Projects created under personal Google accounts outlive the person and are awkward to move.
  • Networking is non-trivialVPC peering, Shared VPC, Private Service Connect and firewall hierarchies are powerful once set up, and painful on the way there.
  • Some products feel strandedApp Engine Standard legacy runtimes and older Firebase products still exist but attract little attention. It is not always clear which path Google wants you on.
  • Service accounts multiplyKeys get generated, emailed, committed to repos and forgotten. A year-old project typically has a dozen service accounts and nobody remembers what half of them do.

Sound familiar?

Common Google Cloud problems we fix

The Google Cloud problems that bring people to us, and what you have once each one is fixed.

container failed to start on PORT=8080

Cloud Run won't come up

The image listens on the wrong port, starts too slowly, or crashes on a missing env var or secret. Logs explorer tells you which, once you know where to look.

Deploys green, revisions tagged

owner: former contractor

The project was inherited

The billing account sits on someone's personal Gmail, the org has no admin, and nobody is sure which services are even in use.

Ownership moved to your org

budget alert: 150% of $500/mo

The bill jumped and nobody knows why

A BigQuery query scanning full tables, a forgotten GKE cluster, or egress from a chatty service. Billing export to BigQuery makes the culprit visible.

Costs attributed and capped

PERMISSION_DENIED on run.services.get

IAM is a knot

Half the team has Owner, the other half can't deploy, and service accounts are scattered across projects.

Roles mapped to real duties

Resources exceeded during query execution

BigQuery jobs fail or cost a fortune

Queries scanning every partition, no clustering, and dashboards hitting raw event tables every minute.

Queries tuned, slots or budgets set

cloudsqlproxy connection refused

Cloud SQL from Cloud Run won't connect

Missing Cloud SQL Client role on the service account, the wrong connector mode, or a private IP without a VPC connector.

App talks to the database cleanly

secret version destroyed

A deploy broke because a secret vanished

Someone cleaned up Secret Manager versions, or a rotation workflow destroyed the one still in use.

Secrets managed with rotation

nodes.available 0

GKE pods won't schedule

Autoscaler blocked by quota, node pool out of a resource, or workloads requesting more CPU than any node can offer.

Cluster right-sized, quotas raised

What we do

Google Cloud help, from rescue to upkeep

Call us for one broken deploy, or hand us the whole org.

Rescue
Deploys failing, bill spiking, or a project you cannot sign into. We take it from there and get you back in control.
IAM and org tidy-up
Moving projects under an org, mapping roles to real jobs, retiring stale service accounts and keys.
Cloud Run and GKE
Containers that start, scale and stop paying attention to themselves. Health checks, concurrency and autoscaling set properly.
Data and BigQuery
Query cost reviews, partitioning and clustering, scheduled transfers and dashboards that don't blow the slot budget.
Cost control
Billing export, budgets with real notifications, committed use discounts where they pay back, and tagged resources.
Managed Tech
IAM reviews each quarter, billing budgets wired to real humans, service account hygiene, and an engineer who already knows your project when a deploy breaks.

How it works

From first message to fixed

  1. Look Free estimate

    Tell us what's going wrong. We look at it, then tell you plainly what's wrong and what it will cost.

  2. Fix

    We fix the cause, not the symptom. If something is down, getting it back up comes first.

  3. Harden

    Backups, monitoring, updates and access, so the same problem can't quietly come back.

  4. Hand back or stay on

    You get a plain-language write-up. Or we stay on for hosting, upkeep and new work.

Often found together

Running Google Cloud alongside something else?

We work on the whole stack, so the problem can be wherever it actually is.

Questions

Google Cloud questions

Something else? Call+1 737 727 7327

Why can't our Cloud Run service start?

Nine times out of ten the container is not listening on the port Cloud Run is probing, or it is crashing at startup on a missing environment variable or secret. The revision logs under the service show the exit reason, and once that is pinned down the fix is usually small.

Can you take over a Google Cloud project from a previous developer?

Yes, and we do this often. We move the project under your organisation, replace the billing account with one you own, audit every service account and role, and hand you back a GCP footprint where the login, the payment method and the admin access are all yours.

Why did our BigQuery bill spike?

Usually a new query or dashboard started scanning full tables instead of partitions, or a scheduled job began running far more often than anyone realised. Turning on billing export to BigQuery and pointing the auditor at the jobs table nearly always finds it inside an hour.

How do we set up IAM properly for a team?

Start at the organisation, not inside one project. Group people by what they do, grant predefined roles to groups rather than individuals, use service accounts for workloads and Workload Identity for anything on GKE, and keep basic Owner and Editor roles out of day-to-day use.

Should we move from App Engine to Cloud Run?

For most new work, yes. Cloud Run gives you the same scale-to-zero convenience without locking your code into a specific runtime, and it costs about the same for typical web workloads. We will tell you plainly if your App Engine app has a reason to stay where it is.

How much does Google Cloud help cost?

Estimates are free. Describe what is broken or unclear and you get a price before any work starts.

Free estimate

Google Cloud got away from you?

Tell us what you're running and what's going wrong. A person reads it and comes back with an estimate.

  • No charge to find out what it will take
  • If it's urgent, say so

Replies come from a person. By sending, you authorize us to email you back. No marketing.