When developing applications for

War. You may find some difficulty in introducing stateful behavior into a stateless environment when you need to connect to a persistent data store. Let’s take a closer look at how to effectively leverage serverless capabilities while avoiding common pitfalls. Common Serverless Disadvantages Regardless of the underlying runtime differences, all serverless functions face the same challenges. They are ephemeral deployments. Existing function instances can be closed at any time and new instances can be created without knowledge of previous processing. This is very detrimental to services that need to access non-temporary data stores. For example, consider what would happen if a function was turned off during a transaction.  new functions to be connected to the database. A long-running query or multiple queries made the function call take longer than average. a serverless environment it is always important to assume that such issues can and will occur.


And then returns the result object. In a non-serverless environment this function will not have any impact on performance but in a serverless environment without some pr photo editing servies otection this function can cause severe damage to your application and your wallet. For example if usage of this application increases significantly you may see the database run out of connections very quickly. This can result in slower response times and timeouts that reduce your effective processing speed. To avoid this parallelization problem let's look at three simple configuration changes you can make to your application. The changes are ordered from least impactful and least difficult to most impactful and most difficult. Changing the Client Connection Pool Size Most packages have a way to modify the number of connections a client keeps .



Open to the underlying database called the connection pool. The number of connections in the pool by default may vary but is usually between . If you refer to the above example if you keep connections open per function it will probably be reduced by an order of magnitude even considering the large number of connections. In most cases setting the pool size to the maximum will keep your application running while also guaranteeing The number of connections will never exceed the number of functions running concurrently. If you are still seeing runaway database connections Avoiding Serverless Headaches To illustrate how the problem arises let’s look at a simple example. Below is a simple handler which accepts a query in the database for an item.