Firestore has a quirk that catches people off guard on the Professional Data Engineer exam. When you create a Firestore database, you pick a location. That choice is permanent. You cannot edit it, migrate it, or convert one type to the other from inside the project. If you change your mind later, you spin up a new GCP project, create a new Firestore database, and move the data over yourself. The exam likes to test whether you understood that constraint before you committed to a design.
This article walks through what "location" actually means in Firestore, the difference between a multi-region and a regional configuration, the two multi-region codes you should know by name, and the way the Professional Data Engineer exam frames this decision in scenario questions.
A Firestore location is the geographic footprint Google Cloud uses to physically store and replicate your data. It is not the same thing as a Compute Engine region or a Cloud Storage bucket region, even though some of the codes look similar. Firestore exposes two flavors of location.
nam5 covers the United States, and eur3 covers Europe. A multi-region database keeps serving reads and writes even if one of the constituent regions goes down.us-central1 or europe-west3. Replication still happens across zones inside that region, so a zonal outage does not take you offline, but a full regional outage does.Both options give you strong consistency. The split is about how wide a blast radius the database survives and how far the data has to travel to reach your users.
Multi-region is the right call when you want the highest availability Firestore offers and when your traffic is spread across a wide geography. Google Cloud publishes a 99.999 percent SLA on multi-region Firestore, which is one nine higher than the regional SLA. The trade-off is cost. You pay more per stored gigabyte and per operation because the data is being kept in sync across multiple physical regions.
The exam phrasing usually points you toward multi-region when a scenario mentions any of the following.
If you see those signals, nam5 or eur3 is the answer.
Regional is the right call when your users are concentrated in one area, when your finance team cares about the storage bill, or when you have a data residency requirement that pins the data to a specific country or jurisdiction.
Data residency is the underrated reason on the exam. A multi-region like nam5 spans multiple US regions, which is fine for most American workloads but is not acceptable if a contract or regulation says the data must sit in a single named region. In those cases you pick a regional location such as us-east1 and accept the lower availability ceiling.
The other common regional trigger on the exam is cost sensitivity. Internal tools, prototypes, and workloads with predictable single-region traffic do not need the extra resilience and do not need to pay for it.
Firestore replicates synchronously inside a location. In a regional database, writes are acknowledged after the data is durably stored across multiple zones in that region. In a multi-region database, writes are acknowledged after the data is durably stored across multiple regions inside the multi-region. Either way, by the time your client gets a success response, the write has survived more than one physical failure domain.
That is why Firestore can offer such strong availability numbers without giving up strong consistency. The cost shows up as slightly higher write latency on multi-region compared to regional, because the write has to land in more physical places before it returns.
Scenario questions on the Professional Data Engineer exam tend to give you a short paragraph about an application and ask you to pick a Firestore location or a configuration. The answer almost always follows a simple decision tree.
nam5 for the Americas and eur3 for Europe.That last point is the one the exam loves. If the answer choices include something like "edit the database settings to change the region," that option is a trap. There is no in-place change for Firestore location.
Firestore location is one of those design decisions where the right answer depends entirely on what the workload needs, but the wrong answer is always the same: picking a location, building on top of it, and then discovering later that you actually needed the other one. Spend the few minutes up front to map your users, your availability target, your budget, and your residency requirements to either a multi-region or a regional configuration before you create the database.
My Professional Data Engineer course covers Firestore in depth alongside every other data store on the exam, with scenario walkthroughs that mirror the way Google phrases location and replication questions.
