PostgreSQL Reviews

Twice now, I have been involved in the decision by a company to migrate away from MS SQL Server to Postgres. The first time, it was simply a matter of scalability. Once you approach 10 TB of data, managing it in MSSQL becomes problematic. You reach limits on performance, backup/recovery and general maintainability. The second company that I assisted in performing this migration chose Postgres due to the TCO as well as the ability to scale the databases horizontally.

The feature that I find most useful (and in fact critical) is the extensibility of Postgres. We installed the extensions that were important to us and ignored anything that wasn’t useful. This allows us to maintain a highly customized configuration that is still able to be supported and maintained by third-party vendors.

One of the key ways that Postgres has improved the functioning of our organization is by freeing up financial resources that can then be applied to upgrading existing infrastructure. A side benefit, of course, is that by bringing in another platform, we have given current staff the ability to grow their skill set and experiment with a new, feature-rich environment. This improves employee satisfaction and makes our CFO happy at the same time.

I would really like to see a more mainstream approach to support what we see as critical extensions. One example is the FDW (foreign data wrapper) for MSSQL. This extension hasn’t been updated in several releases and would benefit from an overhaul. In general, the Postgres community is not as enthusiastic about supporting integration with Windows products (MSSQL, AD, etc.) as they are about other products like Oracle, GIS and full-text searching.

I have been involved in various aspects of Postgres for approximately two years. This includes both single-node installations as well as multi-node clusters using PostgresXL.

The only issue I have ever come up against is internal support. Implementing Linux and Postgres in an environment where only Microsoft has lived has been challenging at times. Administering Postgres on Ubuntu (or any other variant) takes a far different skill set than supporting SQL Server on Windows.

As far as scalability goes, I have yet to identify the limits of Postgres. We will be looking at the newer multi-node options from 2nd Quadrant later this year.

Like any open source product, your mileage may vary. There are several VERY good third-party options for technical support. That being said, this product is not for the faint of heart or the technically unsophisticated shop.

See my answer above. We evaluated both open-source as well as proprietary solutions. Of the open-source solutions we examined, Postgres has the best track record for innovation and enhancements. While the user base is smaller than some of the more established solutions, the fact that it has been able to avoid being “acquired” by a major player is, in my opinion, a plus.

Postgres will work straight out of the box on most platforms. However like all of the database vendors in the Unix space, the ability to modify the configurations are extensive. The degree of complexity is less than Oracle or Sybase but certainly more complex than something like SQL Server. The most important thing to keep in mind is that you must understand how your operating system handles memory. The most complex part of the Postgres installation is, by far, security. I would recommend getting help before tackling the HBA configuration file.

Both times I have been involved in an initial Postgres implementation, we have handled it in-house. It isn’t too hard to implement but you do need some base tech skills including Unix. I would not recommend trying to implement it on a Windows server.

For us, the ROI was almost immediate. We saved several $100k in license costs alone. Overall, the manpower costs to support Postgres and Linux will depend on whether those skills already exist in your enterprise. If you plan to take a Postgres system live in production, I strongly encourage you to look into commercial support.

If you can, do it!

I have used it in the past for some web applications and back-end databases. In my current organization, we are using Microsoft SQL Server.

It is very useful for both structured and unstructured data. You can store unstructured and structured data in PostgreSQL. It is easy to use. You can easily manage things through PostgreSQL Admin.

It is cost-effective. Its on-premise version is free. It is agnostic of on-premise or cloud. You can install it on the cloud or on-premises. It is available with all clouds, and you can also install it on desktop or Windows Servers.

It would be good to have machine learning functionality in this solution, similar to Microsoft SQL Server and other solutions. Machine learning capability for a basic level or a common user would be useful.

It can also have good reporting capabilities.

I have been using this solution for a couple of years.

PostgreSQL has been in the market for a long time. It is quite stable.

It is scalable. In my past organization, its usage had increased a lot. I had implemented data management and many other things on PostgreSQL.

In terms of the number of users, we had hundreds of users who used this solution. For development, we had seven or eight developers. We also had technical support and application teams.

I have not interacted with the support of Postgres because when it is on the cloud, it is managed by the respective cloud provider's team.

We used to provide service to various clients, and we were also providing internal services. We used different solutions in parallel, such as Amazon Redshift, MySQL. MySQL is also free. I have also used Oracle and IBM Db2 in other organizations.

Its installation is simple and easy. If it is in the cloud, you have to go for a subscription. On a desktop, you can install it with normal Unix commands.

I have not done full server version installation myself. If we go for Azure Cloud, its API is available. It takes five minutes to get it up and running on the cloud version. For desktop deployment, you can complete your setup within half an hour. 

It is open-source. If you use it on-premise, it is free. It also has enterprise or commercial versions. If you go for the cloud version, there will be a cost, but it is lower than Oracle or Microsoft. 

I would definitely recommend this solution. It is a very good database to have. It is also very good as compared to other tools.

I would rate PostgreSQL a nine out of ten. 

I use PostgreSQL on-premises to store monitoring data collected by  Zabbix Server.

I wanted a database engine that could handle an ingress of a thousand real-time values per second, delete old items without affecting performance, and handle hundreds of user queries at all times.

The solution had to support high compression and time series data while maintaining data integrity and performance.

I wanted the database engine to be easy to tune, secure, and set up.

PostgreSQL matched those requirements and has regular updates and plenty of official and community support resources.

PostgreSQL greatly improved our monitoring solutions data storage, performance, compression, and processing. Our monitoring solutions run efficiently with little maintenance.

The availability, stability, and reliability of our monitoring solutions greatly improved because the database engine scales out well, is easy to tune, easy to upgrade and manage, and supports extensions and plugins for specific use cases. One such plugin is TimescaleDB and it has proved greatly beneficial for time-series data storage and automatic partitioning of the database.The upgrade of the database has been great too, from 12 to 13 to version 14.

The most valuable feature is support for the Timescale DB extension. We managed to reduce the storage space needed to 10% of the original size, without affecting data integrity, and we significantly improved the performance.

The database engine is easy to manage, the tuning is friendly, and the integration with supported extensions is friendly too.

The database engine is free and open-source, too. Since we did everything internally, it has greatly reduced the costs of setting up our systems.

It also supports diverse kinds of replication, which is crucial for a high availability environment that we plan to set in the near future.

PostgreSQL uses high memory compared to its counterparts when a highly demanding workload with many database connections is in use, especially one that makes many concurrent connections to the database. 

Like many other databases, the tuning is manual through a configuration file. It would be useful if the database engine could detect the specifications of the machine in which it is installed and so bring some levels of auto-tuning. 

PostgreSQL replication support isn't so straightforward for multi-sources and master replicas. It will be great if native support of those replication modes become available in the future.

I have been using PostgreSQL for more than four years.

Stability-wise, I have a great impression.

It can be easily scaled.

We haven't used the official support but judging from the available resources on the website and other outlets it seems their support is good.

Positive

I used other database management systems (MySQL and its variant MariaDB) for my NMS applications before moving to PostgreSQL. I had some optimization issues on MySQL and MariaDB and decided to switch to PostgreSQL, mainly for the TimescaleDB extension support provided on PostgreSQL and which my application natively support including automatic database partitioning and compression. TimescaleDB proved to be helpful since I mostly deal with time series data and the TimescaleDB hypertables improved my applications perfomance greatly.

The initial setup was straightforward, although it needed time to get everything well-tuned. 

I implemented in-house.

The ROI is 100%.

PostgreSQL is free and open-source, so if capable admins are available then the setup cost can be negligible. We use internal resources, so it was completely free for us. One can choose the available official support too.

I evaluated other options including MySQL and its variant MariaDB & Percona Server for MySQL, Oracle DB, and SQLite.

For anybody who is considering this solution, my advice is that it is better to do enough research on the specific database engine requirements.

I highly recommend PostgreSQL with TimescaleDB extension for time-series data.

This solution is very good for getting data. It is very useful for reporting purposes. We also use one of the tools of IBM Cognos for reporting purposes.

For small-scale setups, the solution is stable.

The solution is open-source and free to use.

Currently, we are working with Postgre's economy, and we are not able to implement real-time solutions with our existing architecture. There's a general lack of real-time data from Postgre.

The solution isn't as stable for larger data sets. 

The scalability is limited.

We'd like the solution to be faster. 

I've been working with Postgre for the last two years. However, the organization that I was working with from the beginning has been using this. This was open-source and they've been working with it since 2017.

The product is stable only for working on relatively small data. It's not ideal for large amounts of data. Our business has grown rapidly. We have more customers and our data has grown very rapidly. We need to manage performance and tuning and may need a grander product. The performance could be better.

We're struggling with scalability. It's one of the issues we're facing right now.

This is hosted by this AWS. We generally raise tickets to them whenever we're in need of assistance.

For example, if we're doing any replication or any of that thing, we just raise tickets, and it gets resolved through them.

The support has been pretty good.

I was not involved in the actual deployment of the solution and therefore cannot speak to how difficult or simple the process was.

Currently, I've been given the role to do some research and switch to a different database.

The solution is open-source. We don't need to have a license in order to use it. 

We're currently looking for a more modern solution to replace Postgre. I'm aware of AWS Redshift, Query, and Nextly. These are major players in the market.

We are just a customer of Postgre.

Since the data volume has increased rapidly, we are concerned Postgre won't be suitable for our long-term requirements. We're currently looking for an all-in-one option.

I'd advise users that are not looking for speed or do not have huge amounts of data to try this solution. 

I'd rate the solution at an eight out of ten.

Typically, our team runs the database and then the applications, on Postgres. However, I'm not part of the development process.

The solution is quite stable and very reliable.

On the cloud side of the product, the solution scales quite well.

The initial setup is quick and easy.

I don't work directly on development, however, I haven't heard of any complaints from the development team in general. I can't speak to any features that may be missing. Our team seems quite satisfied with it overall.

It would be great if the solution offered even more integration capabilities.

We've been using the solution for the past two or three years at this point.

The stability, overall, has been good. I have not heard of issues with bugs or glitches. I cannot recall it crashing or freezing. Its performance has been reliable.

The product can scale well. However, this is mostly the case on the cloud, which more easily can scale as there are no physical limitations to hold a company back. If a company needs to scale using this solution, it can do so with relative ease, specifically if they use a cloud deployment.

We only have about ten to 15 or so users on the solution right now. On the products we produce there may be more, however, that varies. There could be a hundred or so users.

We have plans to continue to use the solution going forward.

We don't really use technical support too often. We have our own team that we can turn to, and they can handle most, if not all, issues.

We've used a few other products previously. We're also using MongoDB, or at least, we will be, in an application that we've just started.

The initial setup is not complex at all. I would describe it as straightforward and rather simple. 

The deployment is quick as well. It might have taken us about 45 minutes to an hour or so to get everything up and running. It's great.

We have a technical team or two or three people that can manage the deployment and maintenance. You don't need a big team.

We handled the implementation process ourselves. We didn't need the assistance of any consultants or integrators. It was pretty straightforward, and therefore we didn't need the extra help.

The product is a direct service, and it is free to use. There isn't a licensing fee.

We are simply a customer and end-user. We don't have a business relationship with PostgreSQL.

The solution is deployed both on the cloud and on-premises. We use more than one deployment model.

I can't recall the exact version number we are using, however, it's my understanding that it is not necessarily the latest version.

I'd recommend this product to other organizations. It's worked well for us so far.

In general, I would rate the solution at an eight out of ten.

The database for the accounting systems is our primary use case for this solution. Business software is 80% of my use case in the accounting software. It's a good tool for planning. I work with financial data and the data mining aspect is very important, as I need to access the data easily. The data is the most important thing. Sometimes the software developers forget this part because they just want to save the data on the database, but they don't know how to manipulate the data. The performance of the solution is dependant on how easily and quickly you can obtain access to the data.

I can use Unity Kinematic. It's not easy but you can create a multi-tenant database in Postgres and I think it's a must-have in software development. With the database, you can provide a multi-component at the same service with the same performance, scalability, or all those things. 

PostgreSQL is limited to 32 TB in maximum table size for version 9.6 and earlier. This can be problematic for large databases. We currently don't have the multi-tenant feature and we require that with the contract integrated reassurance. To create have one tenant or more tenants, I need to implement or develop a software solution to connect with the multi-tenant, and to connect each tenant. PostgreSQL is not great for software developers, but it is adequate. For developers, databases such as Firebase and MongoDB technology are better. With PostgreSQL, all you need the database for is to save your data. This function on its own is not enough. The database and applications can become very slow. To debug you need to make the database query display information on your device. Some developers are not experienced in developing databases. This is especially difficult for junior software developers, and they often need more experience to be able to deal with that aspect.

We have been using Postgres for about 10 years. 

It is a stable database.

Because PostgreSQL is limited in regards to the amount of data it can hold in a single table, you often think it might be better to change the technology. Other than that, it is a good solution for business.

Sometimes it is not simple to resolve issues, but the PostgreSQL community sometimes helps to find answers to questions.

I think it is difficult to setup. I amended my ID and Oracle advised me to change it, for it to work.

If you want one tenant, you can use MySQL. If you want a multi-tenant, I think Postgres is better. SQL on PostgreSQL is linked to Oracle. Oracle and Postgres is the same thing, the same language. You will not have a big change for the software developers to migrate from Oracle to Postgres. 

I would rate PostgreSQL 8 out of 10. 

One use case is that we installed it and built it for a customer so that he has access to the database, and he can create whatever he would like to create on it. 

We have deployed it on-premises and in the cloud.

It is easy to install and easy to manage. There is no license on it, so it is free. 

There is high compatibility with Oracle, and there are many tools for the migration of data from Oracle to Postgre. 

It still needs to be more mature and have some backup feature. We are normally dealing with Oracle's data, and we have very strong online tools to back up the data and do other things. PostgreSQL still needs to do more in this area as well as in the high availability area. There are many external tools that you can use for PostgreSQL's high availability, but there is no embedded tool within PostgreSQL for high availability.

It could have a feature similar to Oracle for working on a distributed system. It can have some scripts to improve the monitoring and some tools to do performance analysis. We have a workaround for most of such requirements except for the support for a distributed system, which is very difficult to have. This area should be included in the core of the database itself. 

I have been using this solution for almost one year.

I didn't have any issues, but I think Oracle is more stable.

I didn't have experience with that because we didn't make any changes since we built it. All I have is one server, and I can only have one standby, nothing more. 

We didn't contact them. We didn't face any serious issues that required support.

I am also using Oracle Database. The main difference is the scalability. PostgreSQL could be used for small to medium databases but not for a huge production database. I still prefer to have Oracle Database.

The initial setup was straightforward. It did not take too long. It took maybe one hour to do the installation.

It is free. There is no license on it.

Go ahead and implement it. It is a nice product, but keep a backup and try to use it for small to medium projects or companies. Some of the customers are demanding PostgreSQL nowadays, so we will keep on implementing it.

I would rate PostgreSQL an eight out of ten.

We use the solution for media purposes. We manage the sites of one of the largest sports business papers and multiple TV channels. So, we develop the websites.

We are increasingly using their support for JSON, which we find to be very complete, something I made use of in the past. The solution provides complete support in terms of the SQL dialect and behaves nicely when it comes to transactions. One can change the database structure transactionally. This is one of the few databases that allows this. I like it. 

The solution is comparable in sophistication with that of Oracle. Each product has a few things less and more than the other. We also like that the solution is open source. We have good performance with a small footprint. It's rather nice. It's very robust.

The solution could be improved through an upgrade to the latest version. 

A better graphic user-interface would be nice to see. 

There is nothing I particularly dislike about the solution. The data propagation in master-slave configurations would be a good example. This is one of the features that I understood the least, yet we have it working and use it to propagate from the content management system database to the multiple publishing databases. 

This said, I would like this propagation feature to be simplified for new users and to come with better explanation. However, I will refrain from giving criticism on this point, as I do not know if they already handled this in the last version. Overall, I have only praise for the solution.

I cannot point to anything in particular that we are missing out on at the moment. What comes to mind are features that I have yet to try, although I don't  have any wish lists for PostgreSQL at the moment. I don't know how it stacks up when it comes to the importing and exporting of data. For databases involving this, we just make use of Redshift, which is verified from PostgreSQL and developed by Amazon. 

I have been using PostgreSQL for a couple of years.

The solution is very stable. We have databases that have been running for years.

We have scaled the solution just to three slave machines, but it works well when it comes to master and slave. 

Mostly, support consists of the community and there are several consultancies, should the need arise, although we have never had use of these. There have been no issues on this end. Community support has always been sufficient for us. 

On a single machine, the initial setup is extremely easy. I also find configuration to be very simple. It is similar to MySQL in that a person must know what he wants when it comes to setting up the solution, in which certain features would come into play. Setting it up in a different way will involve the use of multiple search engines. 

With the solution, one installs it and whatever he tunes is optional. Of course, he would have to play with the configurations if he wishes to have specific personality, such as cluster configurations, or tuning for very demanding performance. Overall, for small things or development, one need only install it, start it and it's done. 

While the length of the deployment varies with the configuration, a simple one will take a couple of minutes. 

We deploy the solution both on-premises and on AWS. 

I had my doubts about the functionality before joining this company, as it seemed very complex. It turns out that the solution is actually very simple to set up and we have it working all the time without any problems. It survives the network partitions, so we like this very much. 

My advice is that a person just try it and use it. For me, it beats out JSON and is superior to MongoDB. It works in a completely different way. But, overall, I would rather use PostgreSQL when it comes to starting and manipulating JSON and it boasts superior integrity and performance. Of course, there are specific things that MongoDB does differently. A person's mileage may vary, depending on what he wishes to accomplish. 

I rate PostgreSQL as a nine out of ten and I choose to knock it down a point only because it could use a better graphic user-interface.