How can this bash script be made shorter and better? Is there a way to combine multiple curl statements and if statements in bash? - linux

Question

Status_code_1=$(curl -o /dev/null -s -w "%{http_code}\n" -H "Authorization: token abc123" https://url_1.com)
Status_code_2=$(curl -o /dev/null -s -w "%{http_code}\n" -H "Authorization: token abc123" https://url_2.com)
if [ $Status_code_1 == "200" ]; then
echo "url_1 is running successfully"
else
echo "Error at url_1. Status code:" $Status_code_1
fi
if [ $Status_code_2 == "200" ]; then
echo "url_2 is running successfully"
else
echo "Error Error at url_2. Status code:" $Status_code_2
fi
The main script is scheduled and runs everyday and prints the success message everytime. If the status code is anything other than 200, $Status_code_1 or $Status_code_2, whichever is down prints the error code.
The code is working fine but I want to know how can it be made shorter. Can the curl command from first 2 lines be combined because they have same authorization and everything, it's just the url at the end is different. Also later if statements are pretty much same, only that I'm running them separately for different urls.
Is it possible to write first 2 lines in one line and same for both if statements? I know AND and OR can be used for if statements but say we have 5 urls and 2 are down, how it will print the name of those 2 urls in that case?

Answer 1

Is there a way to combine multiple curl statements and if statements in bash?
curl can retrieve multiple URLs in one run, but then you're left to parse its output into per-URL pieces. Since you want to report on the response for each URL, it is probably to your advantage to run curl separately for each.
But you can make the script less repetitive by writing a shell function that performs one curl run and reports on the results:
test_url() {
local status=$(curl -o /dev/null -s -w "%{http_code}\n" -H "Authorization: token abc123" "$1")
if [ "$status" = 200 ]; then
echo "$2 is running successfully"
else
echo "Error at $2. Status code: $status"
fi
}
Then running it for a given URL is a one-liner:
test_url https://url_1.com url1_1
test_url https://url_2.com url1_2
Altogether, that's also about the same length as the original, but this is the break-even point on length. Each specific URL to test requires only one line, as opposed to six in your version. Also, if you want to change any of the details of the lookup or status reporting then you can do it in one place for all.

Answer 2

To avoid repetition, you can encapsulate code you need to reuse in a function.
httpresult () {
curl -o /dev/null -s -w "%{http_code}\n" -H "Authorization: token abc123" "$#"
}
check200 () {
local status=$(httpresult "$1")
if [ "$status" = "200" ]; then
echo "$0: ${2-$1} is running successfully" >&2
else
echo "$0: error at ${2-$1}. Status code: $status" >&2
fi
}
check200 "https://url_1.com/" "url_1"
check200 "https://url_2.com/" "url_2"
Splitting httpresult to a separate function isn't really necessary, but perhaps useful both as a demonstration of a more modular design, and as something you might reuse in other scripts too.
I changed the formatting of the status message to include the name of the script in the message, and to print diagnostics to standard error instead of standard output, in accordance with common best practices.
The check200 function accepts a URL and optionally a human-readable label to use in the diagnostic messages; if you omit it, they will simply contain the URL, too. It wasn't clear from your question whether the labels are important and useful.
Notice that the standard comparison operator in [ ... ] is =, not == (though Bash will accept both).