These are chat archives for getredash/redash

2nd
Sep 2016
HenroRitchie
@HenroRitchie
Sep 02 2016 09:57
Hi All. What is the best way to display the current date and time on a dashboard?
Arik Fraimovich
@arikfr
Sep 02 2016 10:19
@HenroRitchie if you allow scripts in user input, then you can add some javascript in a textbox widget that shows the current time/date.
HenroRitchie
@HenroRitchie
Sep 02 2016 10:27
@arikfr So the 'text' in the text box can be the script?
@arikfr I have also tried to approach it from another angle by enabling the python scripting and retrieving both the date and time with a python script and returning a table with the time in one column and the date in another. However using the counter visualization nothing gets displayed
Arik Fraimovich
@arikfr
Sep 02 2016 10:34

@HenroRitchie the text box takes markdown and html, so you can use a <script> tag there, as long as you enable scripts in user input (but it's a security risk, and therefore disabled by default).

If using a query is good enough for you, you can use any DB (even Redash's own Postgres) to return the current date & time.

Counter visualization: yes, currently it breaks with non numeric input, but it will be great to fix this.

HenroRitchie
@HenroRitchie
Sep 02 2016 10:36
Aaa, ok, yes that is the problem, because word cloud works fine
Arik Fraimovich
@arikfr
Sep 02 2016 10:37
I think it breaks because it tries to do some numbers comparison .
HenroRitchie
@HenroRitchie
Sep 02 2016 10:38
@arikfr Just to report back on the redis issue. So redash no longer returns a red box with an error. However once a day something goes wrong with redis where it no longer accepts any connection
I have not found the solution to that yet
Arik Fraimovich
@arikfr
Sep 02 2016 10:38
I'm not sure what's better then :worried:
I have to admit that I never had neither of this issues :-\
HenroRitchie
@HenroRitchie
Sep 02 2016 10:40
Me to, I have clients looking at me with watery eyes... I have deployed a bitnami ami image with redash 0.10 to see if it is more stable on aws
Arik Fraimovich
@arikfr
Sep 02 2016 10:41
I doubt it will be more stable and it's harder to support due to their own folders structure and software (like Apache instead of Nginx, which was causing some issues in the past).
Redis usually is very stable strange that it will stop accepting connections at all.
HenroRitchie
@HenroRitchie
Sep 02 2016 10:45
Agreed. I would stick to the normal version, however to find answer I am investigating all avenues. Maybe I should redeploy mine to see if I broken something. When will you release a new version?
Arik Fraimovich
@arikfr
Sep 02 2016 10:46
it will take a few more weeks until 0.12 declared as stable, but it's quite stable except for a bug in the schema browser.
HenroRitchie
@HenroRitchie
Sep 02 2016 14:34
@arikfr When running redis-server --test-memory redis dies
Gavin Todes
@GAV1N
Sep 02 2016 14:34
I'm having problems with an alert that uses the "greater than" operation. I set the reference value to 1500000 and the UI and query clearly show that the current value for that field is 400000, yet for some reason the alert gets triggered. Is anybody else able to successfully use the "greater than" op?
I'm on version 0.11.1+b2095 btw
Arik Fraimovich
@arikfr
Sep 02 2016 14:39
@GAV1N maybe the value in the column isn't really a number but a string?
Gavin Todes
@GAV1N
Sep 02 2016 14:39
It should be - I'm using MySQL's CAST(<field> AS UNSIGNED)
Arik Fraimovich
@arikfr
Sep 02 2016 14:40
hmm.. can you share the json result output of this query? (/api/queries/<id>/results.json)
Gavin Todes
@GAV1N
Sep 02 2016 14:42
Ah that confirms your suspicion :) Here's the relevant part of the json:
{"Target Price": "", "Referral ID": null, "Contacts URL": "", "Referral Name": "", "Highest Target Price This Hour": "400000", "Deal ID": null, "Added At": "", "Stage": "", "Lowest Target Price This Hour": "260000", "Deal Type": ""}
Arik Fraimovich
@arikfr
Sep 02 2016 14:42
:)
Gavin Todes
@GAV1N
Sep 02 2016 14:44
Thanks @arikfr
Now to figure out how to cast it correctly...
Arik Fraimovich
@arikfr
Sep 02 2016 14:44
strange that cast didn't do it :-\
can you try casting it to integer instead of unsigned?
Gavin Todes
@GAV1N
Sep 02 2016 14:52
Ah shoot, I know what I did. The query is a UNION of two queries where the first returns one row with a value in that field cast as an UNSIGNED integer, and the second returns a number of rows with no value in that field (it's a hacky way to have 1 row that shows the max and min value across all records, and then below it all the detail rows with for all records, so that the alert can link to a detail query). Anyway, in the second query, I was setting the value to an empty string instead of NULL.
Thanks for helping, @arikfr
Arik Fraimovich
@arikfr
Sep 02 2016 14:53
:+1:
HenroRitchie
@HenroRitchie
Sep 02 2016 16:23
@arikfr How big is the swap space for the AWS image?
J C Lawrence
@clearclaw
Sep 02 2016 19:08
Is there a way for a task in a DAG to cause the entire DAG to be re-run? I'm specifically looking at cases where a task late in the DAG determines that there was actually a failure in some early task that was previously undetected....and would likely heal if everything were just done again.