Replacing a MongoDB with a Lambda

The project that I currently work on uses a number of microservices. Several of these microservices use MongoDB. The platform provider that we are using to host these databases (compose.io) has announced that they are going to stop supporting Mongo in about two months time.

This has resulted in a thorough investigation of the mongo databases that we are using.

Several of the services use the following architecture:

This has three instances of an application that periodically poll a website before performing sending the result to rabbitmq. They use the mongodb as a lock so that only one of the three perform the operation.

The alternative design was to add an external scheduler to trigger the nodes. The load balancer would pick one of the nodes to perform the work. One of my colleagues pointed out that we could use a lambda to do this. Lambdas are (typically) javascript functions that are cloud hosted.

It it is possible to deploy a lambda with a schedule of how frequently it should be invoked, either with a given frequency or via a CRON-like specification. The lambdas invoke methods on the nodes that asynchronously trigger the process.

Here is a diagram of the replacement:

This has allowed us to replace three MongoDB’s in each of four environments with a much simpler lambda function.

In addition we have been able to remove the mongodb database driver and the internal cron scheduling components. This greatly simplifies these services. There is now less to maintain (less PR’s from dependabot). It is also far easier to test code that is called from an endpoint than one from a scheduled event.

We do still need to replace the remaining MongoDB’s which contain real data but this is a much more manageable job.

We also learned that AWS still uses Node 8 LTS for Lambda.

This is a great starting point for serverless functions.

Announcing About

Last weekend I wrote and released the Elixir package called “about”.

This adds About topics to the iex environment.

It is installed in the usual way:

Add {:about, "~> 0.0.1"} to your deps list in mix.exs

Run mix deps.get

Run iex -S mix

You can now use help topics like:

iex(1)> h About.pipeline

I am looking for suggestions/contributions to the project.

It’s hosted on https://github.com/chriseyre2000/about and available on hex.pm and hexdocs.

Designing Elixir Systems With OTP – Part Three

The B2 edition of this book is out. The new chapter is dedicated to Supervisors. These are correctly described as Lifecycle components.

This is the first time that I have seen the new DynamicSupervisors being used.

Minor note the parameter in

Lifecycle/lib/mastery/boundary/quiz_session.ex for select_session/1 has changed to name from session.

Similarly for answer_question/2

The code samples now clearly state where the code is to be added.

There was a minor typo and some missing code that breaks the final iex session.

Current code can be found at: https://github.com:chriseyre2000/designing_elixir

Elixir iex documentation : C

This is the third part of an exploration of the Elixir base libraries

Here is an introduction to what you get out of the box with iex:

LetterModuleDescription
CCalendarDate, time and calendar functions
CaseClauseErrorThis is a struct used for errors.
CodeUtilities for managing code compilation, code
evaluation, and code loading.
CollectableProtocol to traverse data structures.
Used for putting data into things.
CompileErrorThis is a struct used for errors.
CondClauseErrorThis is a struct used for errors.

Doing the same with :c gives the erlang modules

LetterModuleDescription
ccErlang compiler.
calendarDate, time and calendar functions
cerlCore Erlang abstract syntax trees.
cerl_clausesUtility functions for Core Erlang
case/receive clauses.
cerl_inline???
cerl_setsSet functions.
cerl_treesBasic functions on Core Erlang abstract syntax trees.
codeUtilities for managing code compilation, code
evaluation, and code loading.
code_serverProcess used for live code reloading
compileErlang compiler interface
core_lib???
core_lint
core_parse
core_pp
core_scan
countersLibrary of mutable counters with no locking.

Elixir iex documentation : B

This is the second part of an exploration of the Elixir base libraries

Here is an introduction to what you get out of the box with iex:

LetterModuleDescription
BBadArityErrorThis is a struct used for errors.
BadBooleanErrorThis is a struct used for errors.
BadFunctionErrorThis is a struct used for errors.
BadMapErrorThis is a struct used for errors.
BadStructErrorThis is a struct used for errors.
BaseEncoding and decoding functions
for 16, 32, 64, hex, url
BehaviourMechanism for handling behaviours
(deprecated)
BitwisePerforms bit calculations.

Doing the same with :b gives the erlang modules

LetterModuleDescription
bbase64Base 64 encode and decode
beam_aUsed after code generation before
optimisation. Normalises the code.
beam_asmAssembler for the beam
beam_blockErlang compiler
beam_bsErlang compiler
beam_bsmErlang compiler
beam_cleanErlang compiler
beam_deadErlang compiler
beam_dictErlang compiler
beam_disamErlang compiler
beam_exceptErlang compiler
beam_flattenErlang compiler
beam_jumpErlang compiler
beam_libErlang compiler
beam_listingErlang compiler
beam_opcodesErlang compiler
beam_peepErlang compiler
beam_receiveErlang compiler
beam_recordErlang compiler
beam_reorderErlang compiler
beam_splitErlang compiler
beam_trimErlang compiler
beam_typeErlang compiler
beam_utilsErlang compiler
beam_validatorErlang compiler
beam_zErlang compiler
binaryHandles binary data.
This is the Erlang string type

Elixir iex documentation : A

I recently found out how powerful the iex tool is for documentation.

This is the general form of getting detailed documentation:

This is the start of a long series about the modules that you get out of the box.
1> h Module.function_name/arity

This returns the documentation on this specific function in the specific module.

I also found that typing a capital letter and tab expands the known modules.

Here is an introduction to what you get out of the box with iex:

LetterModuleDescription
AAccessKey based access to structures.
AgentSimple abstraction around state
ApplicationWorking with applications and their callbacks
ArgumentErrorThis is a struct used for errors.
ArithmeticError This is a struct used for errors.
AtomConvenience functions for working with atoms.

Translating Elixir/Erlang terms:

An application is equivalent of a windows DLL, something that you compose a system out of. A typical application would be a logging system.

Atoms are a non-garbage collected resource. They make convenient aliases to global things.

Doing the same with :a gives the erlang modules

LetterModuleDescription
aapplicationGeneric OTP Application
application_controllerKernel supplied application controller
application_masterKernel supplied application master.
Responsible for knowing the topmost
supervisor
application_starterResponsible for starting applications in
a defined order
arrayArray structure, fixed or extensible
atomicsProvides atomic operations using
hardware instructions
authErlang network authentication server
(deprecated)

Working through “Property Based Testing in PropEr with Erlang and Elixir” in Elixir – Part 2

I have returned to this book now that the book is out of beta.

I also started using a different repo: https://github.com/chriseyre2000/property_testing

The only difference between my code and the samples is that I am using the pipe syntax where possible. My Execism.IO background in Elixir is also forcing me to make style changes to keep the code in Elixir format.

Part 1 chapter 4 does use the erlang queue library. It would be good to be able to get Elixir documentation on erlang packages.

The part 2 sample code is broken but I am working on fixing in my repo.

It turned out to be easier to download the code from the pragprog site.

Designing Elixir Systems With OTP – Part Two

I have continued working through this beta book and have updated the github repo: github.com/chriseyre2000/designing_elixir with the samples.

So far this book is easier to follow than say Programming Phoenix since it does not keep changing the code that you are working on. The testing code is normally close to where you have typed it, so finding mistakes does not require too much effort. I would recommend using mix test after each code sample has been typed in as fixing errors can be tricky.

Note the only difference between my code and the supplied sample is the use of the british name Maths rather than Math for the sample code. My degree is in Mathematics so I can’t let a typo like that stand!

The main points from the second half of the book are the very clean validation techniques used. I like that they called the process returned by start_link a session.

The book currently stops at the end of chapter 6, so I will have to resume work on this series once it has been updated.

Designing Elixir Systems With OTP – Part One

These are my notes as I am working my way through this book from PragProg. I am reading a beta version of the book so things may change. I’ll be posting this to my github repo: (https://github.com/chriseyre2000/designing_elixir)

The introduction starts by building a simple counter application as the domain and wraps it in a send/receive process with an API.

The code samples here are clean and clearly ordered. The only note that I have found so far is that I needed to delete the autogenerated Count.ex file to avoid a warning.

Takeaway from the intro: Supervisors are about lifecycles.

Takeaway from type introductions: Send the functions to the data!

Chapter 3 typo: response.ex is misspelt (reported).

excoveralls looks to be a better code coverage tool.