Google Firebase's app testing feature for Android, where you can upload your APK, needs improvement because it's not working very well. It's not very easy because you can upload one version of your app and test it with different phones. Google needs to improve it because I tried it multiple times and found that manually testing the app works better and much faster.

Google should find a way to reduce the solution's read and write. For example, if you have a list of 100 clients for only one click, and if they read, it will be 100 reads, which is expensive. We could use a cache method, but it's still expensive. So Google should reduce the cost a bit. Like Supabase, Google Firebase can have SQL models so that we can have relational and non-relational databases at the same time. A relational database could be much cheaper in some cases.

Integrations and extensions offered by the product are quite limited presently. From an improvement perspective, the tool should try to increase the extensions it provides. I feel that Google is trying to improve in the area of the extensions it provides, but I feel that it would take time.

The product must support more complex queries. It is a NoSQL database. It is a collection of documents. Because of its design, the tool can't support some of our complex queries in SQL and other databases.

The product is not that cheap.

I had a problem deleting the user's files, and I had to move to the paid plan to do that. They have added a plugin to allow you to do that, but I've not implemented that yet. Google Firebase should add extra functionality to do global delete and updates across your database.

The initial setup is not exactly straightforward. 

They really need to work on their performance. It is not so good.

Google Firebase doesn't need significant improvements because it provides a complete package for databases, crash analytics, and reports. 

They have discontinued some very valuable features, including the Firebase UI that I used to log in. They are no longer supporting that feature for authentication.

The cloud functionality can always be better. We cannot, for example, use it outside Google. 

There are some features that are only available online. There aren't very many things you can do in your local environment. 

I would prefer Google Firebase to focus more on its integration capabilities since it is an area in the solution with certain shortcomings. Though Google Firebase does a lot of things, there is no clear documentation for code insights. The product is not easily available for a person to understand, so a person has to do a lot of experiments to understand how things happen in the background. Google Firebase should try to simplify its product for others to implement it, but the code insights or functionalities are not available, and there are no explanations available on what it can and cannot do.

The solution's notification system could be much better. Also, they should add automation aspects to it. The authentication capabilities need expansion as well. In addition, they should enable vertical scaling for real-time databases.

They should add a code library to it. Presently, it is very troublesome to develop new code. Also, there needs to be a document for standardized procedures.

The notifications must be made a little bit better. Firebase Cloud Messaging doesn’t support non-dismissable notifications. I encountered some difficulties migrating the code from Firebase Cloud Storage to another DB.

In Google Firebase, reporting is an area with shortcomings where improvements are required to be made in the future. Sometimes, it's not easy to retrieve data based on the pre-defined reports. There are some things you would like to measure, and it gets really hard to gather from the reports in Google Firebase, so there is a need to make available a custom report.

The solution's pricing could be better.

There's always room for improvement in every product, and the major improvement that's needed in Google Firebase is its simplification, particularly for unsavvy users or non-technical users. If you're not tech-savvy, it would be very difficult for you to use the functions of the solution and understand its security rules.

If you want even non-technical people to find it easy to use Google Firebase, it should be simplified because the difference between screen-derived management and code-derived management isn't complex, so simplification wouldn't be an issue.

Another area for improvement in the solution is the functionality of querying a database, for example, the database in my company is DocumentDB, but Google Firebase doesn't support SQL queries, and using SQL language will dramatically increase usage. Each database has its language for querying, and I like SQL which is a standard query language. It's really difficult to query databases because, on each database, you have to learn a different language, so in DocumentDB and some other databases, there's a layer that gives you the ability to query with regular SQL language, but Google Firebase doesn't have that layer or functionality.

An additional feature I'd like to see in the next version of the solution is better querying. Another feature I'd like added to Google Firebase is for the security to be derived by screens, rather than by code, meaning that it can be automated by screens, and not by coding, then you can take it out of the development area, so it can be maintained by non-tech savvy users.

The platform's bandwidth is quite expensive. Its price could be better.

One of the challenges I was facing when I was using it was the library called Google Dialogflow which allows you to connect your chatbot with an intent. So there is so much adding when you are trying to connect it with your service because you have to do a lot of repetition writing a new intent, putting it step by step so it is good to escape. I believe it should be flexible and dedicate the intents based on the relating and to write a lesser code when doing that.

I would like to have a facility for testing possible scenarios by writing dummy data and then doing different queries. This would help me to better develop applications.

The only thing that I won't do on Firebase is the hosting side of things for a progressive web app because it feels like you actually still need to download the app first before you can actually get into the functionality of the app.

So, it's not really a website, but it's web-based, and the app needs to be like a downloadable app. In that sense, I'm using normal Linux hosting and PHP for normal websites. I still have to split my software into a web app. If there's something that needs to be improved, it is that if you publish an app as web-based that you don't have to physically download the app or that the app doesn't download in the background. 

I would rather not publish as a progressive web app. Initially, on first opening up, the app seems very slow. After that it's quick, but it feels like the app needs to be downloaded first before you can actually get into the functionality of the app.

I think the migration into a data warehouse, or Google BigQuery, should be a bit more automated. There are still some manual hiccups, and it wasn't as user friendly as some of the other components.

I would like to see faster data migration and some support for other languages in the next release. Google Firebase has functions, but it'll help to do a .NET Core kind of function or a bit of additional support for other languages. Right now, I feel like it's very Google-centric.

Firebase's real-time information isn't technically real-time. It takes a while to upload user activity into the app. It's not as fast as Datadog, the other solution we use. 

In my humble opinion, Google Cloud Functions (which is still a BETA version) needs to have more control over the uploaded functions. That is, it needs more options for controlling the functions itself. I might think about some characteristics that can improve a lot this:

- A START button to start the function. This button should be
incorporated in each one of the functions that the developer created and
deployed in Google Cloud Functions;
- A STOP button to stop the function. This button should be incorporated
in each one of the functions that the developer created and
deployed in Google Cloud Functions;
- A RESTART button following the save behavior as above buttons with
RESTART capability.
- A better and improved log view. The actual LOG view has a Slow
performance to show the log items and also it doesn't allow fast searching
for past log items. For example, whenever we need to look for a block of
log items that happened in the past, it's very difficult to access that
block of code.
- Capability to save the log to an external tool.

We had some compatibility issues with Elixir. It had some library compatibility issues. We did a workaround through Google Cloud Functions and achieved the desired output.

We have experienced many challenges with the realtime database. Every time we need to retrieve chat history that has multiple observers, it will execute multiple times. The UI for the realtime database could be simplified.

Google Firebase could improve the ability to create our own reports, other than what's available on the dashboard. For example, if we want to customize the report or export the data, this is a large task. 

Crashlytics was not very helpful, it doesn't give enough information to narrow down the problem.

Earlier I was using Firebase, however, now I am using Firestore. Firestore is much better than Firebase as it is the advanced version of Firebase and also it reduces the developer's time even more. It provides for queries. In Firebase you don't have that many queries to implement, get, or sort data. Firestore provides a lot of queries including MySQL. We find it much better. Firebase should offer some more of the features that exist in Firestore.

The solution needs to offer capabilities to work offline. Like Firestore, we should be able to do insertion, updating, and deletion offline. 

When we connect to the internet, it should automatically update data to the server.

As we were not able to calculate the amount of traffic and the bandwidth, we were not able to calculate the pricing accurately. There was no way of knowing how the rise in traffic would impact the price.

The cold start-up of the cloud function needs to speed up. For me personally, I'm not experienced in anything there, however, I've just read online it's a problem.

Their technical support needs to be enhanced with more support for customers.

They should also improve the monitoring and control for the database to monitor the performance and fine-tuning of the database itself. Automatic detection of potential errors or failures in the database would also be very helpful. For example, I would like to receive notifications for the admin to take appropriate action for database failures.

The solution could improve by providing an improved SDK tool and automatic indexing for databases.

We use Firebase for media processing and we are growing more. As we grow, the pricing goes up and it's high compared to competitors. The solution is expensive. 

In the next release, they should focus more on realtime features. 

Firebase uses JSON, and the storage format is very different to that of SQL. This does not allow for easy migrations.