Alternatives to Knative

How Uber developed the open source, end-to-end distributed tracing Jaeger , now a CNCF project:

Distributed tracing is quickly becoming a must-have component in the tools that organizations use to monitor their complex, microservice-based architectures. At Uber, our open source distributed tracing system Jaeger saw large-scale internal adoption throughout 2016, integrated into hundreds of microservices and now recording thousands of traces every second.

Here is the story of how we got here, from investigating off-the-shelf solutions like Zipkin, to why we switched from pull to push architecture, and how distributed tracing will continue to evolve:

https://eng.uber.com/distributed-tracing/

(GitHub Pages : https://www.jaegertracing.io/, GitHub: https://github.com/jaegertracing/jaeger)

Bindings/Operator: Python Java Node.js Go C++ Kubernetes JavaScript OpenShift C# Apache Spark

To provide employees with the critical need of interactive querying, we’ve worked with Presto, an open-source distributed SQL query engine, over the years. Operating Presto at Pinterest’s scale has involved resolving quite a few challenges like, supporting deeply nested and huge thrift schemas, slow/ bad worker detection and remediation, auto-scaling cluster, graceful cluster shutdown and impersonation support for ldap authenticator.

Our infrastructure is built on top of Amazon EC2 and we leverage Amazon S3 for storing our data. This separates compute and storage layers, and allows multiple compute clusters to share the S3 data.

We have hundreds of petabytes of data and tens of thousands of Apache Hive tables. Our Presto clusters are comprised of a fleet of 450 r4.8xl EC2 instances. Presto clusters together have over 100 TBs of memory and 14K vcpu cores. Within Pinterest, we have close to more than 1,000 monthly active users (out of total 1,600+ Pinterest employees) using Presto, who run about 400K queries on these clusters per month.

Each query submitted to Presto cluster is logged to a Kafka topic via Singer. Singer is a logging agent built at Pinterest and we talked about it in a previous post. Each query is logged when it is submitted and when it finishes. When a Presto cluster crashes, we will have query submitted events without corresponding query finished events. These events enable us to capture the effect of cluster crashes over time.

Each Presto cluster at Pinterest has workers on a mix of dedicated AWS EC2 instances and Kubernetes pods. Kubernetes platform provides us with the capability to add and remove workers from a Presto cluster very quickly. The best-case latency on bringing up a new worker on Kubernetes is less than a minute. However, when the Kubernetes cluster itself is out of resources and needs to scale up, it can take up to ten minutes. Some other advantages of deploying on Kubernetes platform is that our Presto deployment becomes agnostic of cloud vendor, instance types, OS, etc.

#BigData #AWS #DataScience #DataEngineering

Currently been using an older version of OpenFaaS, but the new version now requires payment for things we did on the older version. Been looking for alternatives to OpenFaas that have Kafka integrations, and scale to 0 capabilities.

looked at Apache OpenWhisk, but we run on RKE2, and my initial install of Openwhisk appears to be too out of date to support RKE2 and missing images from docker.io. So now looking at Knative. What are your thoughts? We need support to be able to process functions about 10k a min, which can vary on time of execution, between ms and mins. So looking for horizontal scaling that can be controlled by other metrics, than just cpu and ram utilization, but more so, for example if the wait is over 5 scale out.. Issue with older openfaas, was scaling on RKE2 was not working great, for example, I could get it to scale from 5 to 20 pods, but only 12 of them would ever have data, but my backlog would have 100k's of files waiting.. So even though it scaled up, it was as if the distribution of work was only being married to specific pods. If I killed the pods that had no work, they come up again with no work, if I killed one with work, then another pod would scale up and another pod would start to get work. And On occasion with hours, it would reset down to the original deployment allotment of pods, and never scale up again, until I go into Kubernetes and tell it to add more pods.

So hoping to find a solution that doesn't require as much triage, to work with scaling, as points in time we are at higher volume and other points of time could be no volume.

CodeFactor being a #SAAS product, our goal was to run on a cloud-native infrastructure since day one. We wanted to stay product focused, rather than having to work on the infrastructure that supports the application. We needed a cloud-hosting provider that would be reliable, economical and most efficient for our product.

CodeFactor.io aims to provide an automated and frictionless code review service for software developers. That requires agility, instant provisioning, autoscaling, security, availability and compliance management features. We looked at the top three #IAAS providers that take up the majority of market share: Amazon's Amazon EC2 , Microsoft's Microsoft Azure, and Google Compute Engine.

AWS has been available since 2006 and has developed the most extensive services ant tools variety at a massive scale. Azure and GCP are about half the AWS age, but also satisfied our technical requirements.

It is worth noting that even though all three providers support Docker containerization services, GCP has the most robust offering due to their investments in Kubernetes. Also, if you are a Microsoft shop, and develop in .NET - Visual Studio Azure shines at integration there and all your existing .NET code works seamlessly on Azure. All three providers have serverless computing offerings (AWS Lambda, Azure Functions, and Google Cloud Functions). Additionally, all three providers have machine learning tools, but GCP appears to be the most developer-friendly, intuitive and complete when it comes to #Machinelearning and #AI.

The prices between providers are competitive across the board. For our requirements, AWS would have been the most expensive, GCP the least expensive and Azure was in the middle. Plus, if you #Autoscale frequently with large deltas, note that Azure and GCP have per minute billing, where AWS bills you per hour. We also applied for the #Startup programs with all three providers, and this is where Azure shined. While AWS and GCP for startups would have covered us for about one year of infrastructure costs, Azure Sponsorship would cover about two years of CodeFactor's hosting costs. Moreover, Azure Team was terrific - I felt that they wanted to work with us where for AWS and GCP we were just another startup.

In summary, we were leaning towards GCP. GCP's advantages in containerization, automation toolset, #Devops mindset, and pricing were the driving factors there. Nevertheless, we could not say no to Azure's financial incentives and a strong sense of partnership and support throughout the process.

Bottom line is, IAAS offerings with AWS, Azure, and GCP are evolving fast. At CodeFactor, we aim to be platform agnostic where it is practical and retain the flexibility to cherry-pick the best products across providers.

AWS Lambda Serverless Amazon CloudWatch Azure Functions Google Cloud Functions Node.js

In the last year or so, I moved all Checkly monitoring workloads to AWS Lambda. Here are some stats:

• We run three core functions in all AWS regions. They handle API checks, browser checks and setup / teardown scripts. Check our docs to find out what that means.
• All functions are hooked up to SNS topics but can also be triggered directly through AWS SDK calls.
• The busiest function is a plumbing function that forwards data to our database. It is invoked anywhere between 7000 and 10.000 times per hour with an average duration of about 179 ms.
• We run separate dev and test versions of each function in each region.

Moving all this to AWS Lambda took some work and considerations. The blog post linked below goes into the following topics:

• Why Lambda is an almost perfect match for SaaS. Especially when you're small.
• Why I don't use a "big" framework around it.
• Why distributed background jobs triggered by queues are Lambda's raison d'être.
• Why monitoring & logging is still an issue.

https://blog.checklyhq.com/how-i-made-aws-lambda-work-for-my-saas/

At my company, we are trying to move away from a monolith into microservices led architecture. We are now stuck with a problem to establish a communication mechanism between microservices. Since, we are planning to use service meshes and something like Dapr/Istio, we are not sure on how to split services between the two. Service meshes offer Traffic Routing or Splitting whereas, Dapr can offer state management and service-service invocation. At the same time both of them provide mLTS, Metrics, Resiliency and tracing. How to choose who should offer what?

As for the new support of service mesh pattern by Kong, I wonder how does it compare to Istio?

Recently we were looking at a few robust and cost-effective ways of replicating the data that resides in our production MongoDB to a PostgreSQL database for data warehousing and business intelligence.

We set ourselves the following criteria for the optimal tool that would do this job: - The data replication must be near real-time, yet it should NOT impact the production database - The data replication must be horizontally scalable (based on the load), asynchronous & crash-resilient

Based on the above criteria, we selected the following tools to perform the end to end data replication:

We chose MongoDB Stitch for picking up the changes in the source database. It is the serverless platform from MongoDB. One of the services offered by MongoDB Stitch is Stitch Triggers. Using stitch triggers, you can execute a serverless function (in Node.js) in real time in response to changes in the database. When there are a lot of database changes, Stitch automatically "feeds forward" these changes through an asynchronous queue.

We chose Amazon SQS as the pipe / message backbone for communicating the changes from MongoDB to our own replication service. Interestingly enough, MongoDB stitch offers integration with AWS services.

In the Node.js function, we wrote minimal functionality to communicate the database changes (insert / update / delete / replace) to Amazon SQS.

Next we wrote a minimal micro-service in Python to listen to the message events on SQS, pickup the data payload & mirror the DB changes on to the target Data warehouse. We implemented source data to target data translation by modelling target table structures through SQLAlchemy . We deployed this micro-service as AWS Lambda with Zappa. With Zappa, deploying your services as event-driven & horizontally scalable Lambda service is dumb-easy.

In the end, we got to implement a highly scalable near realtime Change Data Replication service that "works" and deployed to production in a matter of few days!

Heroku Docker GitHub Node.js hapi Vue.js AWS Lambda Amazon S3 PostgreSQL Knex.js Checkly is a fairly young company and we're still working hard to find the correct mix of product features, price and audience.

We are focussed on tech B2B, but I always wanted to serve solo developers too. So I decided to make a $7 plan.

Why $7? Simply put, it seems to be a sweet spot for tech companies: Heroku, Docker, Github, Appoptics (Librato) all offer $7 plans. They must have done a ton of research into this, so why not piggy back that and try it out.

Enough biz talk, onto tech. The challenges were:

• Slice of a portion of the functionality so a $7 plan is still profitable. We call this the "plan limits"
• Update API and back end services to handle and enforce plan limits.
• Update the UI to kindly state plan limits are in effect on some part of the UI.
• Update the pricing page to reflect all changes.
• Keep the actual processing backend, storage and API's as untouched as possible.

In essence, we went from strictly volume based pricing to value based pricing. Here come the technical steps & decisions we made to get there.

1. We updated our PostgreSQL schema so plans now have an array of "features". These are string constants that represent feature toggles.
2. The Vue.js frontend reads these from the vuex store on login.
3. Based on these values, the UI has simple v-if statements to either just show the feature or show a friendly "please upgrade" button.
4. The hapi API has a hook on each relevant API endpoint that checks whether a user's plan has the feature enabled, or not.

Side note: We offer 10 SMS messages per month on the developer plan. However, we were not actually counting how many people were sending. We had to update our alerting daemon (that runs on Heroku and triggers SMS messages via AWS SNS) to actually bump a counter.

What we build is basically feature-toggling based on plan features. It is very extensible for future additions. Our scheduling and storage backend that actually runs users' monitoring requests (AWS Lambda) and stores the results (S3 and Postgres) has no knowledge of all of this and remained unchanged.

Hope this helps anyone building out their SaaS and is in a similar situation.