commit
Diff
tech2019.md | 204 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
images/borg.png | 0
images/red.jpg | 0
images/uber.jpg | 0
4 files changed, 204 insertions(+)
@@ -1,0 +1,204 @@
---
---
# <!--fit--> Tech 2019
_- nemo_
---
# <!-- fit -->2-feet-rule
---
# Caveats 🤷
- Non-Exhaustive
- Aligned to Razorpay
---
# For every Tech
1. What is it?
2. Why should you care?
3. What can we use it for?
---
<!--  -->
<!--  -->
---
# serverless
## What is it?
If your application isn't called frequently enough, your computing resources can be brought up Just In Time.
---
## Why should you care?
- Code runs instantly
- You pay for what you run
- Scaling is easier
---
## Sample Usecases
- Any kind of Batch Jobs that take <15 minutes
- A rarely visited website
- If-this-then-that (Infrequent Trigger based Actions)
- When a user uploads a file ->
- When a user sends an email ->
---
## Limitations
- Vendor-Lockins
- Not all Languages
---
# blockchain
---
# what is it?
Techno-Libertarians's dream of removing central authorities from the world's economy, to only end up crashing it entirely.
---
# what is it?
blockchain is a verifiable distributed database that doesn't require trusted parties
---
---
---
---
---
# why should you care?
---
# really, should I care?
If you trust developers to write code that isn't buggy, and software developers to regulate finance better than the government does - Maybe?
---
---
# the google hype cycle
- Map Reduce -> Apache Hadoop
- BigTable -> Apache HBase
- Chubby -> Apache Zookeeper
- Spanner -> Cockroach DB
- Borg -> kubernetes
---
# Borg
---
# What is it?
Large scale cluster management tool to help you run thousands of applications
---
# why should you care?
- we run kubernetes and all of our services run on kubernetes
- kubernetes is commoditization of the cloud (GCP/AWS :fist_raised:)
- k8s advantage :tm:
---
# <!--fit-->logs v/s metrics v/s tracing
---
# logs
- easiest
- most verbose
- no bird eye view
---
# metrics
- most important
- dimensions/cardinality of a metric
---
# what metrics
---
# USE
> For every resource, check utilization, saturation, and errors.
- utilization: the average time that the resource was busy servicing work [2]
- saturation: the degree to which the resource has extra work which it can't service, often queued
- errors: the count of error events
---
# tracing
- hardest
- needs to be done well
---
---
# <!-- fit --> microservices
---
# microservices
aka: How to Trade Development complexity for Operational complexity
Binary files /dev/null and a/images/borg.png differ
Binary files /dev/null and a/images/red.jpg differ
Binary files /dev/null and a/images/uber.jpg differ