I am getting increasingly frustrated with the quality of the packages in NPM.
Is seems that it’s ok to randomly change code and leave things in a broken state (or at least have a stack of errors for simple usage.
Author: chriseyre2000
Dev-ops or No-ops
The team that I am working on has been described by our outgoing contractors as a DevOps team. We develop, deploy and support the site. We add monitoring to everything we can (sometimes after it goes wrong once). We are an agile team loosly working in a scrum setup. The trick we have is to always dedicate one pair (or half a pair if the team is odd numbered that week) to the operations channel. The team member that is on the support rota gets first option of working on the operations tickets this week.
Dev-Ops is normally about having a dedicated team that perform automation and monitoring of systems. We do have that happening but from within the team. There is no passing it to an “other” team. In that sense we have achieved No-Ops – there literally is no ops team that we can pass off problems too (OK we have various vendors to manage but that is a different matter).
Heroku Process Types
I have been experimenting with the Heroku scheduler.
This means creating a custom process types.
We found a fun exception: You can’t put hypens in the name.
It does not fail with any kind of error, it just fails to install.
Combining d3 and Angular
Here is an example of using Angular to add d3 as a directive:
http://odiseo.net/angularjs/proper-use-of-d3-js-with-angular-directives
Gradle on Google Cloud
This looks like the obvious library to try using google cloud with:
Google Bigquery
I have been experimenting with using google Bigquery recently.
It looks to be an incredibly useful (and cheap) means of storing and using lots of data.
In fact it is so cheap that asking management for permission in a meeting can cost more in staff time than a years use of the data. This means that with the right policy it is possible to store whatever data you think might be of use.
The downside is the documentation. It looks comprehensive at first until you try to use it for something. The java code samples are full of sections such as:
// Insert useful code here
There does not even seem to be a comprehensive list of the datatypes so here goes:
STRING
BOOLEAN
TIMESTAMP
INTEGER
FLOAT
These fill in for where there should be key points (such as defining the fields in a Bigquery table).
Bigquery does have a sensible structure:
Projects are for billing.
These contain datasets that are used for access control (we have a set of these per environment : qa, demo).
Datasets contain tables (and views – but I have not used these yet).
Liquibase command line details
http://www.liquibase.org/documentation/command_line.html
This could provide an easy means of building a tool to check the validity of a liquibase migration script (it would be good to use this as a preview).
I would recommend using the xml version of the migration scripts. YAML is too odd and has so little tool support. We lost an hours development time by a misplaced space in an index file…
How to use S3 as a maven repository
This is a useful article on using S3 as a private maven repository:
http://jmchung.github.io/blog/2015/03/01/using-amazon-s3-as-a-private-maven-repository/
It is better to use this than try to use a git repository – git repos make inefficient mavern repos.
Vodafone are Idiots
They will send a phone to an address that is not associated with the bank details that they have!
Vodafone have just fallen victim to a classic scam.
A courier delivers a phone to an address (unexpectedly). Half an hour later a motorcyclist collects the “mistaken delivery”.
The delivery company is genuine, the phone is real yet no actual bank details have been checked. Vodafone have sent a phone out without checking the address associated with the bank account that is paying for it.
In addition the customer services automated phone system requires you to have an account number before they will talk to you.
The online chat system (clearly staffed from India) is more useful and they are able to provide details of the fraud team number – but even that maze has a few dead ends. You get to choose 1 for fraud or other phone problems, but the next menu can’t cope with general fraud.
Neo4j is incredibly useful
I have been a fan of Neo4j for a while. Until recently I had only been using it locally on my machine to investigate a few simple uses.
Recently I have found about the heroku addon graphenedb (https://elements.heroku.com/addons/graphenedb). This is a relatively cheap (for small graphs it is free) cloud installation of Neo4j.
This meant with a little effort I could have a set of scripts extract the current state of my employers CMS system and then provide an upto date view of the data in a trivially queriable form.
I have found that the easiest way to get control over the graph is to simply start from scratch each day. The script extracts a set of nodes and a list of relationships between them and loads it up. I know that it would be much faster to use the csv import tools but I have just to find a satisfactory means of getting the labels to be set correctly.
